[{"content":"I am happy to write letters of recommendation for students whom I know well enough to recommend honestly and enthusiastically. The type of recommendation I can provide depends on the purpose of the application and the extent of our academic interaction.\nGraduate and professional schools expect much more detailed evaluations than summer internships, so the level of familiarity needed varies accordingly.\nBefore You Ask Before requesting a recommendation, please ask yourself the following questions.\nQuestion ✓ Have I given at least four weeks\u0026rsquo; notice? ☐ Have I taken at least one course with Professor Kmetko? ☐ Does our level of interaction match the type of application I am pursuing? ☐ Have I prepared my résumé or CV and any other required materials? ☐ Have I provided the deadline and submission instructions? ☐ If you can answer \u0026ldquo;Yes\u0026rdquo; to these questions, I will usually have everything I need to evaluate your request promptly.\nWhat to Include with Your Request Please request a recommendation at least four weeks before the deadline, whenever possible.\nWhen requesting a letter, please include:\nyour résumé or CV, a list of all classes you have taken with me and semesters (e.g. phys 41 in Fall 2023), a brief description of the opportunity, your personal statement, if applicable, the deadline, and submission instructions. What Type of Recommendation Can I Request? Application Minimum Interaction Summer internship — short-term academic or industry opportunities Completed at least one course with me. Transfer to another college — transfer admission Completed at least one course with me and demonstrated consistent engagement in class. Employment — full-time, part-time, or campus employment Completed one or more courses with me and demonstrated professionalism, reliability, or initiative. Graduate school — M.S. or Ph.D. programs Significant interaction beyond coursework, such as multiple courses, an independent project, research, or serving as my PALS. Medical, dental, or other professional schools Extensive interaction that allows me to comment on your academic ability, character, maturity, professionalism, and potential for success. Building a Strong Recommendation The strongest letters come from students whom I know well. You can help me write a stronger recommendation by:\nparticipating actively in class, visiting office hours, discussing your academic or career goals, completing an independent or course project, taking more than one class with me, or serving as my PALS or participating in the TF program. When I May Decline Occasionally, I may decline a request if:\nI do not know you well enough to write a strong recommendation; the request is made too close to the deadline; or I cannot honestly provide an enthusiastic recommendation. If I decline, it is not intended as a negative judgment. Rather, I believe you will be better served by asking someone who can write a stronger and more detailed letter.\nA Final Note A recommendation letter is most valuable when it contains specific examples rather than general praise. The better I know your work, your goals, and your growth as a student, the stronger and more useful my recommendation will be.\n","permalink":"https://jankmetko.fyi/teaching/recommendation-letter-policy/","summary":"Read before you ask.","title":"Recommendation Letter Policy"},{"content":"If you emailed me an explanation for an ordinary absence, I appreciate it.\nI want to make sure the policy is clear: the drop-the-lowest policy is what covers ordinary absences.\nThe drop-the-lowest cannot be “saved” for later because your absence has a legitimate explanation.\nOrdinary situations include things like:\na cold a short illness a routine medical appointment a transportation problem oversleeping a work conflict a family obligation a minor emergency a bad day a mistake in planning These things happen. That is why the drop exists.\nAgain, I appreciate your professionalism in explaining why you were absent. Feel free to stop by my office with any questions.\n","permalink":"https://jankmetko.fyi/teaching/drop-the-lowest-policy/","summary":"A clarification of the drop-the-lowest-score policy.","title":"What the Drop Is For"},{"content":"A Simple Formula for Student Emails A good email has five parts:\nA greeting A brief statement of why you are writing Relevant context, if needed A specific request, if you are making one A closing A professional email is not just polite. It is also clear, specific, and consistent with course policies.\nA Less Effective Email I was sick and missed the reading quiz. Let me know how to make it up. I am attaching my missed homework. Also, can you look over my solutions and show me where I made a mistake?\nWhy This Email Does Not Work Well First, the email is missing a greeting. Always begin with something like:\nHi Jan, Hello Professor, Hello,\nSecond, the email asks for exceptions to course policies. Ordinary absences are already handled by the drop-the-lowest policy. Missed assignments cannot be submitted by email. Email is also not the right place to ask for tutoring or detailed review of homework solutions.\nThird, the email is missing a closing. Always end with something like:\nBest, Thank you, See you soon,\nA Better Version Hi Jan,\nI wanted to let you know that I missed class and will use the drop-the-lowest policy for the missed assignment. I also had difficulty with several homework problems, so I plan to come to office hours to ask questions.\nThank you,\nThis version is better because it is polite, acknowledges the course policy, and identifies the appropriate next step.\nFinal Advice Before sending an email, ask yourself:\nIs there a specific action I am asking for?\nIf the answer is no, you may not need to send the email.\n","permalink":"https://jankmetko.fyi/teaching/how-to-write-effective-emails/","summary":"Write emails that are clear, professional, and useful.","title":"How to Write Professional Emails"},{"content":"Special Studies 49 allows students to complete an independent academic research project under instructor and department supervision. The course counts as elective credit in the department offering it.\nWho Is Eligible? Students should have:\nCompleted at least one semester of college work. Taken at least one course in the department where the independent study will take place. Earned at least a 2.0 (C) GPA in that department or field. Not be on academic probation. Before You Apply You must first develop your own project idea and find an instructor willing to supervise it. The project should relate to their area of expertise.\nYour written proposal should explain:\nWhat you plan to study. The scope and content of the project. At least three learning objectives. How you will complete the project. How your work will be evaluated and graded. Helpful questions to answer:\nWhat do you plan to achieve? How will you accomplish it? When will each part be completed? How will your work be evaluated? Application Steps Obtain a Special Studies 49 application from the sponsoring instructor. Complete the form with your instructor. Deadlines Applications should be submitted at least one week before the deadline to add classes to allow time for revisions and approvals.\nCourse Expectations Students meet regularly with their instructor throughout the semester.\nThe instructor may require:\nExams, including a final exam Research papers Presentations Other assignments related to the project Unit Limits A maximum of 6 units of Special Studies 49 may count toward an AA degree or the 60-unit transfer requirement. 3 units per semester is considered a reasonable workload. Enrolling in more than 18 total units in a semester requires an approved petition. ","permalink":"https://jankmetko.fyi/teaching/howto-independent-study/","summary":"Independent Study: Special Studies 49","title":"How to Get Credit For a Project"},{"content":"Image Files Needed Transfer the light, dark, and flat images from the telescope computer into your directory.\nIf your image file names contain spaces, replace the spaces with underscores. Otherwise, applications may hang unexpectedly during processing on Linux machines.\nIn the terminal, navigate to the directory with your images and type:\nrename \u0026#39;s/\\s/_/g\u0026#39; ./*.* If the rename command is not installed, type:\nsudo apt update sudo apt install rename Information Needed The following information will be needed later:\nsite latitude and longitude telescope focal length telescope aperture size sensor width and height sensor pixel size Some applications will parse this information automatically from the FITS header, while others will expect you to enter it manually.\nTo view the FITS header:\nLaunch Autostar IP. If launching for the first time, go to Options → Preferences and make sure Std. of Dev. is checked under Auto Contrast on Load. Load any one of your light images by clicking the Open icon. Click Tools → Image Information, then press View FITS header. The information in the FITS header has been written by the image acquisition software and should be verified independently for consistency.\nUse a plate-solving algorithm to obtain the plate scale and check that the focal length and sensor pixel size correspond to that plate scale.\nTo verify:\nGo to https://nova.astrometry.net/. Follow the directions to upload any one of your image files to the server. Wait until the status changes to Success, then click Go to results page. Check the plate scale. In the example below, it is 0.664 arcsec/pixel. Next, verify that the astrometry plate scale corresponds to the focal length and pixel size by going to https://astronomy.tools/ and choosing Calculators → CCD Calculators.\nMake sure the pixel size is adjusted by your binning factor, if any.\nIn the example screenshots above, the plate scales match. The pixel size can be trusted, so the plate scale helps confirm that the effective telescope focal length is correct.\nIn this example, ZWO quotes the pixel size for the ASI2600MM camera as 3.76 microns. At 2×2 binning, this gives a pixel size of 7.52 microns. The astrometry result therefore helps confirm that the effective focal length of the telescope is indeed 2350 mm, as reported in the FITS header.\nImage Calibration The light images must be corrected with dark and flat images.\nDark images help subtract thermally generated counts. Flat images correct for geometrical aberrations such as vignetting.\nPreparing the Master Dark Image Launch Autostar IP. Click Group → New. Navigate to your image directory containing the dark images. Use * characters to filter the selection. Otherwise, you will have to ctrl-click the individual images you want. For example, to choose all and only images with 80.00s exposure in the file name, type the * characters as shown below. Then shift-click to choose the filtered selection.\nClick Group → Edit to verify that the correct file selection has been loaded into the group. In the example below, only files with 80.00s are loaded into the group, as desired.\nClick Group → Combine to average the pixel counts across all dark images. Save the resulting file as MasterDark with the default extension. Preparing the Master Flat Image Repeat the same process as with the dark images.\nClick Group → New. Navigate to your image directory containing the flat images. Use * characters to filter the selection. Otherwise, you will have to ctrl-click the individual images you want. For example, to choose all and only images with G-RGB in the file name, type the * characters as shown below. Then shift-click to choose the filtered selection.\nClick Group → Edit to verify that the correct file selection has been loaded into the group. In the example, only files with G-RGB are loaded into the group, as desired.\nClick Group → Combine to average the pixel counts across all flat images. Save the resulting file as MasterFlat with the default extension. Calibrating the Light Images Load the light images into a group, then calibrate them as follows.\nClick Group → New. Navigate to your image directory containing the light images. Use * characters to filter the selection. Otherwise, you will have to ctrl-click the individual images you want. For example, to choose all and only images with G-RGB in the file name, type the * characters as shown below. Then shift-click to choose the filtered selection.\nClick Group → Edit to verify that the correct file selection has been loaded into the group. In the example below, only files with G-RGB are loaded into the group, as desired.\nClick Group → Calibrate. Press the Select buttons to choose your Master Dark and Master Flat files. Check the Include boxes for both files. The calibration operation will automatically save the calibrated files into the directory with the light images, prepending each file name with the Calibrated prefix.\nTo view the calibrated images, click Group → Examine and use the scroll buttons to browse through the calibrated images.\nThe calibration files are now ready to be used in the next step described in SOAR Photometry, Part 2.\n","permalink":"https://jankmetko.fyi/teaching/img-processing-part1/","summary":"Autostar IP instructions","title":"Autostar Image Processing Part 1"},{"content":"These instructions assume that you have calibrated your light images as described in SOAR Photometry, Part 1.\nIn this example, we will perform photometry on the target star at the center of the field using AutoStar Image Processing. Make sure that your calibrated images are still loaded in the Group where you left off in Part 1.\nSelect Group → Align → One Star.\nThe program will step through all the images and display their merged result. The star trails show how the mount drifted or the telescope tube flexed, causing the stars to appear streaked.\nClick OK, then click and drag to draw a box around the single bright streak near the center. As soon as you release the mouse button, the program will align the images.\nAfter the images have been aligned, draw a square around the target star and select Set Log File.\nUse the default log-file name or enter a name of your own.\nNext, draw a square around a reference star whose magnitude is known. You can find its magnitude using an AAVSO plot chart.\nSelect Set Reference Magnitude.\nThe reference star used in this example has a magnitude of 11.3. Enter the magnitude and click OK.\nSelect Group → Photometry.\nClick at the center of the target star. Verify the photometry settings, then click Proceed.\nAfter the program finishes processing the images, you can create the light curve by opening the log file in Excel.\nIn Excel, open the log file and make sure that text files are visible in the file-type menu.\nThe Text Import Wizard will display the contents of the log file. Accept the detected settings and click Finish to load the measurements into Excel.\nThe imported data can now be used to plot the target star\u0026rsquo;s brightness as a function of time.\nAnother method of creating the light curve is described in SOAR Photometry, Part 3.\n","permalink":"https://jankmetko.fyi/teaching/img-processing-part2/","summary":"Autostar IP Instructions","title":"Autostar Image Processing Part 2"},{"content":"This document assumes you have calibrated your light images as described in SOAR Photometry, Part 1.\nIn this example, we obtain the light curve for the variable star SZ Lyn, collected on November 22, 2024.\nLarge Ensemble Photometry Observatory Information Launch LEPhot. Under the Observatory Info tab, verify or modify the information as needed. Consult the FITS header as described in Part 1. If the focal length is uncertain, it can be obtained independently from the plate-scale information using astrometry described in Part 1.\nThe detector width and height can be obtained from the FITS header:\nwidth (mm) = width (pixels) × pixel size (microns) / 1000 After confirming that this information works during initial processing, return to this tab and press Save. Ensemble Photometry Session Click the Ensemble Phot. Session tab. Press New Session. All information is reset. Press Load Ref. Image. Note: You may need to change the file type to All Files in order to see files with the .fits extension.\nChoose any one of your calibrated images (avoid the first or last image in the sequence). Verify that the fields populate with the correct information after loading the reference image.\nImportant: If your image files contain spaces in their filenames, the application may hang during later steps without explanation. Consult Part 1 for instructions on replacing spaces with underscores.\nImportant: The Check Star Name field cannot remain empty. If it is left empty, the application may hang during later steps without explanation. The name itself is arbitrary, but it is best practice to choose a known check star near your object from the AAVSO APASS catalog.\nChoosing the Check Star Go to the AAVSO Variable Star Plotter: https://apps.aavso.org/vsp/\nType the name of your object (in this example, SZ Lyn) or its RA and DEC coordinates. Enter the field scale, preferably matching the size of your image field. The center of the chart will be on your target object. Suitable check stars are labeled with numbers that approximately indicate their magnitudes. For example, the star labeled 109 next to SZ Lyn is approximately magnitude 10.9.\nClick Photometry Table for This Chart. Locate star 109, whose AUID is:\n000-BJR-415 Enter this value into the Check Star Name field in LEPhot.\nMarking the Reference Points Explore the zoom and pan controls if needed.\nPress Mark Sky Bkgnd and click on a region containing only sky background.\nPress Mark Target and click on your target star.\nPress Mark Ck. Star and click on the check star selected from the AAVSO chart.\nRecognizing the check star may take some effort. If necessary:\nregenerate the AAVSO chart with a different scale compare star patterns carefully remember that the chart orientation may not match your image orientation Verify the remaining fields: Apass Filter Name should match your filter. Photometry Type should be set to Differential. Other fields should remain Automatic. Press Plate Solve. The various XY coordinates correspond to the points clicked on the image. The red and green circles indicate the target and check stars.\nTroubleshooting If the application hangs:\nVerify that Check Star Name is not empty. Verify that the reference image filename contains no spaces. If No Solution is found:\nVerify that the focal length agrees with the astrometric plate scale determined in Part 1. Verify that the observatory information is correct. Verify that the sensor width and height are entered correctly in millimeters. Measuring the Images Press Measure Images. In the dialog box, shift-click to select all calibrated images. You may need to change the file type to All to see .fits files.\nIf the application hangs, verify that none of the filenames contain spaces.\nLEPhot uses the known magnitudes of APASS stars within the image and performs a fit to determine the magnitudes of the target and check stars.\nWhen processing is complete, a confirmation dialog will display the number of images measured.\nPress Plot Photometry to display the resulting light curve. Press Save \u0026hellip; to save the light curve in: CSV format Tabbed format AAVSO format Press Save Session. Period Search Press New Fit to clear the fields.\nPress SelectSessions and choose the session saved in the previous section.\nMultiple session files from different observing times may be selected for the same object.\nPress Plot Time Series to display the light curve. Press Search For Fit. Iteratively adjust:\nMinimum Period Maximum Period Search Step The blue curve indicates the fitting error as a function of period.\nThe lowest fit error occurs at approximately:\nPeriod ≈ 2.8 hours Enter this value into the Best Fit field.\nPress Plot Fit. The fit produces a phase plot of the light curve.\nThe repeating data points may appear slightly offset. Fine-tune the Best Fit value manually:\nEnter a new value. Press Plot Fit. Repeat until the phase curve becomes as smooth as possible. The smoothest curve in this example is obtained with:\nPeriod = 2.89 hours You may also perform a polynomial fit, for example of order 2, to produce a smooth model of the light curve. Conclusion For this data set, we conclude that SZ Lyn has a period of approximately 2.89 hours. Continue to Part 4.\n","permalink":"https://jankmetko.fyi/teaching/img-processing-part3/","summary":"Autostar IP instructions","title":"Autostar Image Processing Part 3"},{"content":"This document assumes you have saved the light curve generated with the Large Ensemble Photometry application as a CSV file, as described in Part 3.\nIn this example, we determine the period of the variable star SZ Lyn using the VSTAR application.\nInstall VSTAR Download and install VSTAR from:\nhttps://www.aavso.org/vstar\nPrepare the CSV File Open the CSV file generated in Part 2 and delete all columns except:\nJD Target Mag Save the file again as a CSV containing only these two columns.\nLoad the Data into VSTAR Load the CSV file into VSTAR.\nPerform a Fourier Analysis Click: Analysis → DC DFT standard scan This produces a Data Compensated Discrete Fourier Transform (DC DFT) of the light curve.\nBecause the example series does not have a name, it is labeled as unspecified.\nDetermine the Candidate Frequency Click on the highest (red) data point in the power spectrum to obtain the candidate frequency.\nIn this example, the candidate frequency is:\n8.28 cycles/day The corresponding period is:\nPeriod = 1 day / 8.28 = 0.1208 days = 2.899 hours Conclusion The Fourier analysis performed by VSTAR yields a candidate period of:\nPeriod ≈ 2.899 hours This value is in excellent agreement with the period determined previously using LEPhot, confirming that SZ Lyn has a period of approximately 2.89 hours.\n","permalink":"https://jankmetko.fyi/teaching/img-processing-part4/","summary":"Autostar IP instructions","title":"Autostar Image Processing Part 4"},{"content":"Selecting the right aperture radius requires a balance. The aperture should be large enough to capture most of the star\u0026rsquo;s light but small enough to exclude unnecessary background noise and light from neighboring stars. The goal is to maximize the measurement\u0026rsquo;s signal-to-noise ratio, or SNR.\nA photometric aperture consists of a circular region centered on the star and a surrounding sky annulus used to measure the background.\nAnatomy of a photometric aperture. Source: Soka University of America.\n1. The FWHM Rule of Thumb For most applications, the quickest method is to base the aperture radius on the Full Width at Half Maximum, or FWHM, of the stars in the image.\nMeasure the FWHM of several typical, unsaturated stars in the image. Set the aperture radius to approximately 1.5 to 2.0 times the FWHM. This range usually captures most of the stellar flux while limiting the amount of background noise included in the measurement.\n2. Curve-of-Growth Analysis For higher-precision measurements or images with unusual point-spread functions, you can construct a curve of growth.\nMeasure the target\u0026rsquo;s flux through a series of concentric apertures with increasing radii. Plot the total enclosed signal against the aperture radius. The measured signal rises steeply at first and then begins to level off as the aperture captures the faint outer wings of the star\u0026rsquo;s profile. Choose an aperture near the point where the curve begins to flatten, before the accumulation of background noise begins to degrade the measurement.\nA standard curve of growth. Source: Starlink.\n3. Maximizing the Signal-to-Noise Ratio As the aperture radius increases, the amount of light collected from the star grows more slowly. Background noise, readout noise, and dark-current noise continue to increase as more pixels are included.\nBecause the area of a circular aperture is proportional to the square of its radius, the background noise grows approximately in proportion to the aperture radius.\nIf the aperture is too small:\nsome of the star\u0026rsquo;s flux is excluded, small centering errors have a greater effect, and variations in atmospheric seeing can change the fraction of light inside the aperture. If the aperture is too large:\nadditional sky pixels contribute little or no starlight, those pixels add noise to the measurement, and the resulting SNR decreases. Many photometry programs can calculate the SNR for several aperture radii. For a particular image, the radius that produces the highest SNR is generally the best choice.\nThe Sky Annulus After selecting the aperture radius, define a sky annulus around the star. The annulus is used to measure the local sky background so that it can be subtracted from the stellar signal.\nInner radius Place the inner boundary far enough from the star that the annulus does not include the faint outer wings of its profile. A typical starting point is approximately four to five times the aperture radius or FWHM.\nOuter radius Make the outer boundary large enough to include enough sky pixels for a reliable background measurement. A typical starting point is approximately six to seven times the aperture radius.\nThe annulus should not contain neighboring stars, cosmic-ray hits, image defects, or strong background gradients.\nKeep the Aperture Consistent Once you select an aperture radius, use the same radius for the target and comparison stars within the image.\nA star\u0026rsquo;s profile extends beyond any finite aperture, so every aperture captures only a fraction of its total light. Using the same aperture for every star ensures that approximately the same fraction is measured, preserving the accuracy of relative photometry.\n","permalink":"https://jankmetko.fyi/teaching/selecting-aperture-in-photometry/","summary":"Choose an aperture radius that maximizes photometric accuracy and SNR.","title":"Selecting an Aperture Radius for Astronomical Photometry"},{"content":"Telescope Startup Prepare the Computer Bypass driver-signature enforcement: Open Settings and go to System → Recovery. Next to Advanced startup, click Restart now. After the computer restarts, select Troubleshoot → Advanced options → Startup Settings. Click Restart. When the list of options appears, press 7 to select Disable driver signature enforcement. Turn On the Equipment Connect the two USB cables to the computer. Turn on the power strip. Turn on the mount. Turn on the focuser. Turn on the filter wheel. Start and Connect the Software Launch the SciTech mount controller and unpark the telescope. Launch TCF-S Focus Commander and connect the focuser. Launch PHD2 Guiding and connect the camera and mount. Launch TheSkyX and connect it to the mount. Launch N.I.N.A. Load the correct N.I.N.A. profile through Options. In N.I.N.A., connect: Camera Filter wheel Focuser Mount Guider Start the camera cooler. Align and Focus the Telescope In TheSkyX, slew to a bright target. Caution: If the telescope is too far out of synchronization, use the paddle and Telrad to navigate to the target.\nMove the focuser to its center position, approximately 3500. Use the secondary-mirror screw to focus the telescope rather than the focuser. Take an image in N.I.N.A. Adjust the secondary mirror until the image is focused and centered. Launch SkyView from the SciTech controller. Synchronize the mount. Telescope Operations Use TheSkyX to navigate to the target. Use N.I.N.A.\u0026rsquo;s Framing tool to plate solve and slew and center the target in the middle of the frame. Switch to the required filter. Run AutoFocus. Take a trial exposure. Check the image for saturated pixels. Confirm that the Airy disks of the stars are round and free of aberrations. Set up guiding in PHD2. Take a trial guided exposure in N.I.N.A. Configure the sequencer to acquire the images. Take dark, flat, and bias frames as needed. Telescope Shutdown Disconnect all equipment in N.I.N.A., then close the application. Disconnect all equipment in PHD2, then close the application. Disconnect the telescope in TheSkyX, then close the application. Disconnect the focuser in TCF-S Focus Commander, then close the application. Park the telescope in SciTech, then close the application. Turn off the filter wheel. Turn off the focuser. Turn off the mount. Turn off the power strip. ","permalink":"https://jankmetko.fyi/teaching/cook-dome-checklists/","summary":"Start, operate, and shut down the Cook Dome telescope safely.","title":"Cook Dome 16-Inch Telescope Checklists"},{"content":"These instructions will allow you to browse through galaxies in your region of interest using TOPCAT , for example in Virgo.\nOpen the TAP Query Window.\nSelect the HEASARC service.\nTo search for galaxies whose major axis \u0026gt; 5 in Virgo(within 20 degrees), type the following ADQL text:\nSELECT TOP 10000 * FROM ugc WHERE 1=CONTAINS( POINT(\u0026#39;ICRS\u0026#39;, ra, dec), CIRCLE(\u0026#39;ICRS\u0026#39;, 187.697167, 12.338972, 20) ) AND blue_major_axis \u0026gt; 5 To obtain the RA and DEC coordinates for Virgo, in the main window, open Cone Search from the VO menu.\nType Virgo in Object Name and press Resolve.\nManually copy the RA and Dec into the TAP Query text, then close the VO window.\nRun the query and close the TAP window. Then sort the resulting table by blue_major_axis in descending order.\nNext, open the Actions window.\nCheck Display HIPS cutout and change Size in Pixels to 600. Then close the window.\nNext, plot the galaxy positions on the Sky Plot.\nThe default should display the positions in the equatorial system.\nDouble-click your table to open it. Check that it is ordered by the blue_major_axis column.\nClick on the first row in your table. The action that you defined earlier will automatically trigger the display of the galaxy and also highlight the galaxy in the Sky Plot.\nContinue to browse through your galaxies in Virgo by clicking on the rows in your table, or you can click on a galaxy in the Sky Plot. If you would like to look at a different region of the sky, repeat the instructions in this part, but choose to resolve the coordinates of your new object in the VO.\n","permalink":"https://jankmetko.fyi/teaching/view-galaxies-in-topcat/","summary":"Browse galaxies in TOPCAT using TAP queries and sky plots.","title":"Viewing Galaxies in TOPCAT"},{"content":"These instructions will allow you to browse through spectra of RR Lyrae variable stars in your region of interest using TOPCAT.\nOpen the TAP Query Window.\nSelect the ARI-Gaia Service.\nTo search for the variable stars, specify the RA and DEC of your region center and type in the following ADQL text:\nSELECT * FROM gaiadr3.gaia_source WHERE 1=CONTAINS(POINT(\u0026#39;ICRS\u0026#39;, ra, dec), CIRCLE(\u0026#39;ICRS\u0026#39;, 271.83743686, 9.56384673, 20.0)) AND has_xp_sampled = \u0026#39;True\u0026#39; AND source_id IN ( SELECT source_id FROM gaiadr3.vari_rrlyrae WHERE int_average_g \u0026lt; 13 ) Run Query and close the TAP window.\nNext, plot the variable star positions on the Sky Plot.\nThe default should display the positions in the equatorial system.\nNext, open the Actions window.\nIn the Actions window, check Invoke Service and choose View DataLink Table for the Action.\nDouble click your table to open it.\nWhen you click on any row, the action will automatically launch the DataLink Table. In the DataLink Table, choose Externally calibrated … spectra. For Action, choose Plane Plot Table, and check Auto-Invoke.\nFrom here on, when you click on any row in your Table, the RR Lyrae spectrum will be shown in the Plane Plot and the position of the RR Lyrae will be highlighted on the Sky Plot.\n","permalink":"https://jankmetko.fyi/teaching/view-rrlyrae-spectra-in-topcat/","summary":"Browse RR Lyrae spectra in TOPCAT using Gaia data.","title":"Viewing RR Lyrae Spectra in TOPCAT"},{"content":"These instructions will allow you to construct an H-R diagram of a large sample of stars with absolute magnitudes corrected for distance in TOPCAT. You can also view the spectrum of any selected star and inspect its image to compare its color with its location on the H-R diagram.\nOpen the TAP Query Window.\nSelect the ARI-Gaia service.\nEnter the following ADQL query:\nSELECT designation, source_id, ra, dec, parallax, bp_rp, phot_g_mean_mag, phot_g_mean_mag + 5 * log10(parallax / 100) AS mag_g FROM gaiadr3.gaia_source WHERE parallax \u0026gt; 10 AND parallax_over_error \u0026gt; 10 AND phot_bp_mean_flux_over_error \u0026gt; 10 AND phot_rp_mean_flux_over_error \u0026gt; 10 AND astrometric_excess_noise \u0026lt; 1 AND has_xp_sampled = \u0026#39;true\u0026#39; Run the query and close the TAP window.\nNext, create the H-R diagram using the Plane Plot.\nSet the axes as follows:\nX: bp_rp Y: mag_g Open the Axes tab and check Y Flip so that brighter stars appear near the top of the diagram.\nNext, open the Actions window.\nIn the Actions window:\nEnable Invoke Service. Choose View DataLink Table as the action. Enable Display HIPS cutout. Double-click your table to open it.\nClick on any row in the table. The DataLink Table will open automatically.\nIn the DataLink Table:\nSelect Externally calibrated \u0026hellip; spectra. Set Action to Plane Plot Table. Check Auto-Invoke. From this point on, selecting any star in the table will automatically:\ndisplay its spectrum in the Plane Plot, show an image of the star centered in the image window, highlight the star\u0026rsquo;s position on the H-R diagram. You can also click directly on any star in the H-R diagram to display its spectrum and image.\nAs an exercise, compare:\na bright red star (upper-right region of the H-R diagram), and a bright blue star (upper-left region of the H-R diagram). Notice how their spectra, colors, and locations on the H-R diagram are related.\n","permalink":"https://jankmetko.fyi/teaching/construct-hr-diagram-in-topcat/","summary":"Create an H-R diagram and explore stellar spectra with Gaia DR3.","title":"Constructing an H-R Diagram in TOPCAT"},{"content":"These instructions will allow you to browse through RR Lyrae variable stars in your region of interest using TOPCAT.\nOpen the TAP Query Window.\nSelect the ARI-Gaia service.\nSpecify the RA and Dec of the center of your region of interest, then enter the following ADQL query:\nSELECT g.source_id, g.ra, g.dec, g.pm, g.parallax, r.int_average_g, r.pf, r.peak_to_peak_g FROM gaiadr3.gaia_source AS g INNER JOIN gaiadr3.vari_rrlyrae AS r ON g.source_id = r.source_id WHERE CONTAINS( POINT(\u0026#39;ICRS\u0026#39;, g.ra, g.dec), CIRCLE(\u0026#39;ICRS\u0026#39;, 271.83743686, 9.56384673, 20) ) = 1 AND r.int_average_g \u0026lt; 13 Run the query and close the TAP window. Then sort the resulting table by int_average_g in ascending order.\nNext, plot the variable star positions on the Sky Plot.\nThe default should display the positions in the equatorial coordinate system.\nDouble-click your table to open it. Verify that it is ordered by the int_average_g column.\nThe table lists the RR Lyrae variable stars, with the brightest stars at the top. It also includes each star\u0026rsquo;s oscillation period (pf) and peak-to-peak brightness variation (peak_to_peak_g).\n","permalink":"https://jankmetko.fyi/teaching/view-rrlyrae-in-topcat/","summary":"Locate and examine RR Lyrae variable stars with Gaia DR3.","title":"Viewing RR Lyrae Variables in TOPCAT"},{"content":"Pictures from the MWO SOAR 2026 program.\nPaula is explaining solar physics in the tunnel of the Snow telescope.\nStudents are standing around the Snow telescope spectrograph. The brightly illuminated disk is the projection of the Sun on the slit.\nAnother view of the spectrograph, with Patricia taking a close look.\nPatricia is pointing at the Snow telescope coelostat, an astronomical tracking device with a flat mirror driven by a clock mechanism to follow the Sun.\nSteve is explaining how he draws sunspots on the projection of the Sun at the bottom of the 150-foot solar tower.\nJohn is explaining the operation of the Snow telescope spectrograph.\nPaula is explaining the basics of solar physics.\nPaula is explaining the operation of the solar Lunt telescope in the Snow telescope tunnel.\n","permalink":"https://jankmetko.fyi/personal/soar-2026-photos/","summary":"Photos from the 2026 SOAR program at Mount Wilson Observatory.","title":"MWO SOAR 2026 Photos"},{"content":"Close your eyes for a minute. Now ask yourself—was it a minute?\nSo. You think you know what time is.\nAt first glance, time seems familiar. We measure it with clocks, track it in seconds, and define it using precise quantum standards such as the cesium-133 atomic clock, where one second corresponds to a fixed number of atomic oscillations. That definition gives the impression that time is a well-understood universal backdrop against which everything happens.\nBut a deeper look at physics reveals something unsettling: we are exceptionally good at measuring time, but physics does not clearly tell us what time is.\nConsider a cesium clock. Suppose you observe a simple interval—say, a car moving from your house to the next house. You can measure the duration of that motion by counting cesium oscillations during the interval. But the measurement merely correlates two physical processes: the car\u0026rsquo;s motion and the cesium atom\u0026rsquo;s oscillations. The atom is not producing time; it is a stable process we agree to use as a reference.\nIn physics, the symbol $t$ is not something found inside nature. It is a parameter used to describe change consistently across different systems. The atom does not define time—it only provides a reliable way to label change.\nEinstein\u0026rsquo;s relativity deepens the picture. Relativity removes the idea of a single universal time and replaces it with spacetime, where time depends on motion and gravity. Two observers in different states of motion or gravitational fields will disagree about durations between the same events. Time is no longer a universal flow, but part of a geometric structure.\nYet relativity still does not explain what time is. Relativity describes how time behaves, not what time fundamentally refers to. Even proper time, measured along a worldline by an ideal clock, is a way of parameterizing change rather than deriving time from within the system.\nNow consider a different system: a blob of creamer dropped into hot coffee. The blob spreads by diffusion, and its diameter $D(t)$ increases as the system evolves. In principle, the spreading blob could serve as a clock. We could time the car\u0026rsquo;s journey from your house to the next house by measuring how much the creamer spreads during the interval.\nThe diffusion process works because the blob\u0026rsquo;s diameter changes monotonically and can be calibrated against a cesium clock. But again, the creamer is not defining time. The spreading blob is another physical system whose evolution we map onto the same parameter.\nThe same structure appears throughout physics. Equations describe systems evolving with respect to $t$, and many different physical processes can be used to measure it. But nothing inside those systems explains what $t$ itself refers to. Every clock assumes a parameter that orders change before the clock is chosen.\nThermodynamics does not change the situation. Entropy increase gives a direction, distinguishing past from future, but entropy does not define time itself. The thermodynamic arrow only tells us which direction along the parameter corresponds to increasing disorder.\nPhysics therefore presents a consistent but strange picture. We can build clocks from almost any stable or monotonic process. We can map one clock onto another. We can describe how time behaves under extreme conditions. Nevertheless every description treats time as something already present in the background rather than something derived from the systems being described.\nThe surprising conclusion is not that time is mysterious in a vague philosophical sense, but something more precise: physics provides no final explanation of what time is. Physics provides a framework for describing change, and within that framework, time is the parameter that organizes change.\nIntuition tells us that time is one of the most obvious features of reality. Modern physics suggests something more unsettling. We can measure time with extraordinary precision, we can use almost any regular process as a clock, and we can describe time\u0026rsquo;s behavior across a wide range of systems—but none of those achievements tells us what time is.\nSo. You really do not know what time is.\n","permalink":"https://jankmetko.fyi/personal/what-is-time/","summary":"So. You think you know what time \u003cem\u003eis\u003c/em\u003e.","title":"What Is Time?"},{"content":"At first glance, consciousness seems impossible without time. To remember anything, a mind must compare its present state with something represented as past. Even self-awareness seems temporal: I recognize myself not merely as something that exists now, but as the same being who existed yesterday and years ago. Consciousness appears to have sequence built into it. Here we are going to claim that consciousness may not require time as a fundamental feature of reality. It may require only ordered relations among physical states, which the mind represents from the inside as temporal flow.\nPhysics complicates the assumption that time is fundamental because physics does not explain what time is. We know how to measure time, but measurement is not explanation. A clock tells us how one physical process compares with another; it does not reveal time as a substance inside nature. In equations, time appears as a parameter, $t$. Physics tells us how systems behave with respect to $t$, but it does not clearly tell us what $t$ fundamentally is.\nRelativity deepens the problem. Different observers can disagree about durations depending on their relative motion and gravitational fields; there is no universal “now” shared by everyone. Time is not an absolute background ticking identically for all observers; it is integrated into spacetime geometry. One interpretation of relativity treats past, present, and future as parts of a four-dimensional structure rather than as stages of a flowing present. That interpretation is not the only possible reading of relativity, but it makes the problem vivid: perhaps the psychological perception of a “flowing present” is not a direct reflection of fundamental reality. Perhaps temporal flow is generated by how a conscious system represents its own location within an ordered physical structure.\nIf time is a structural relation among events rather than a metaphysical flow, we can reframe the brain as a physical system. At any local configuration, its neurons, synapses, chemical gradients, and electrical states form a relational structure. Call the whole brain state $S$. In ordinary language, we write $S(t)$, implying that the brain passes through external moments. But if $t$ is stripped of flow, the system can be described as a sequence of lawful relations among configurations:\n$$ S_1 \\rightarrow S_2 \\rightarrow S_3 \\rightarrow \\cdots $$\nor, more abstractly,\n$$ S_{n+1} = F(S_n) $$\nHere, $F$ represents the lawful relation by which one configuration gives rise to, constrains, or permits another. If the laws are deterministic, $F(S_n)$ may identify a unique successor. If the laws are not deterministic, $F(S_n)$ may identify a set of possible successors. Either way, the ordering of brain states need not depend on a flowing substance called time. “Time” may be the name we give to our internal experience of structural ordering.\nThe relational account reframes the connection between consciousness and time. The mind does not require access to a metaphysical temporal flow; it requires states that stand in asymmetric, ordered relations to other states. Memory can be understood as a state containing traces of lawful predecessors. Anticipation can be understood as a state modeling possible successors. Self-awareness can be understood as a system representing itself as located within a causal pathway. The conscious system does not need to find time as an object inside nature. It needs to organize relations among states in a way that can be experienced from within.\nThermodynamics provides one example of physical asymmetry. In closed macroscopic systems, entropy tends to increase, defining the thermodynamic arrow of time. If $H(S)$ represents the entropy of a state, we often find that:\n$$ H(F(S_1)) \u0026gt; H(S_1) $$\nEntropy increase can distinguish one direction of a relational chain from the other. But entropy alone cannot explain consciousness. Entropy tells us why many physical processes have an arrow, but it does not explain why an ordered process should be experienced from the inside. The brain also complicates a simple entropy story. It is an open system fed by energy, nutrients, and sensory input. Its local neural complexity increases during development while the larger environment pays the thermodynamic cost.\nSo the ordering principle relevant to consciousness cannot be a simple measure of disorder. A better idea is that a conscious state is defined by its position within a causal history:\n$$ H_n = (S_0, S_1, \\ldots, S_n) $$\nOn that view, a conscious state is not merely an instantaneous physical arrangement. It is a state whose meaning depends on the pathway by which it came to occupy its present role. A memory has meaning because the current state contains traces of earlier states. A desire has meaning because the current state models possible successor states. A self-model has meaning because the system represents itself as persisting through its own ordered history. The state is not an isolated point; it is embedded in a causal structure.\nThe causal-history model helps explain why introspective self-awareness develops gradually. A baby may experience hunger, warmth, pain, comfort, and perception. But a full self-model requires accumulated concepts, language, bodily maps, social expectations, memories, and patterns of self-recognition. Only then can the brain organize its own informational trajectory into a coherent, persisting “self.” The self may not be present first, waiting to collect experiences. The self may emerge as the brain becomes capable of representing its own ordered development.\nThe causal-history model also helps explain the privacy of experience. Even if Alice and Bob are performing identical actions in the present, their internal experiences remain distinct because Alice’s state $S_n$ encodes a different informational lineage than Bob’s. Alice’s present state contains traces of Alice’s body, memories, associations, fears, habits, language, and past choices. Bob’s present state contains traces of Bob’s. Bob cannot directly feel what it is like to be Alice because Alice’s experience is not an isolated object floating at an instant. It is a state whose meaning is derived from its location within her causal graph.\nTraced backward, that graph has no clean boundary. Alice’s biological organization extends through her parents, evolutionary history, chemistry, stellar formation, and ultimately back toward the earliest state of the universe. In principle, everything belongs to one universal causal web. In practice, a subject is individuated by the portion of that history that becomes locally encoded, functional, and self-referential. The self does not emerge from an absolute first moment. It emerges where a causal trajectory becomes sophisticated enough to represent its own ordering.\nThe causal-history picture does not solve the hard problem of consciousness. It does not explain why physical relations should produce subjective experience at all. But it changes the form of the question. Instead of asking how consciousness can exist in time, we can ask how a system embedded in lawful relations constructs the experience of time. The mystery shifts from “How does consciousness move through time?” to “How do ordered physical relations become represented as temporal flow from the inside?”\nThe feeling of temporal flow may be analogous to the feeling of color. The physical universe contains wavelengths of light, but consciousness constructs the experience of redness. Similarly, the physical universe may contain ordered relations between states, while consciousness constructs the experience of past, present, and future. The universe contains structure; consciousness experiences structure as flow. Time is not built into consciousness because time is a fundamental substance of reality. Time is built into consciousness because time is what ordered, informational relations feel like when represented from the inside.\n","permalink":"https://jankmetko.fyi/personal/consciousness-without-time/","summary":"What if time is something consciousness invents?","title":"Consciousness Without Time"},{"content":"Have you ever noticed how relational mathematics and physics are? In mathematics, calculus can be rebuilt from limits, limits from real numbers, real numbers from set-theoretic constructions, and those constructions from logic, membership, and axioms. In physics, objects appear relational in a similar way. A tree depends on wood molecules, wood molecules depend on atoms, atoms depend on electrons and nuclei, nuclei depend on quarks bound by fields, and in modern physics particles themselves are often understood as excitations or roles within quantum fields. What about consciousness? If we keep following these chains of dependence, we seem to expect a bottom on which everything else rests. But perhaps the lesson is stranger: perhaps the bottom is not a thing at all. Maybe relations, rather than things, are at the core of reality.\nA set-axiomatic model can represent the universe as a structure of relations rather than as a collection of self-standing objects. However, no such model can serve as an exhaustive, closed description of reality if it is consistent, axiomatizable, and powerful enough to express arithmetic. The path to that conclusion begins with the empty set, because the empty set reveals how mathematics can build definite structures from minimal formal assumptions. The same question then returns in physics: perhaps the natural universe, too, is not made of independent things but of relations that make things appear. A complete account of reality must eventually face not only particles, fields, and laws, but also the subject for whom anything appears at all. Consciousness therefore enters the argument not as a side-topic, but as the hardest test of formal description. Subjective experience appears undeniably true from the first-person point of view, but may resist complete derivation from within the formal syntax of a physical model.\nBegin with the empty set, written as $\\varnothing$ or $\\lbrace\\rbrace$. Its cardinality is zero because it contains no elements. However the empty set can still serve as an element of another set, so $\\lbrace\\varnothing\\rbrace$ has cardinality one. In the von Neumann construction, that distinction becomes the basis of counting: $0=\\varnothing$, $1=\\lbrace\\varnothing\\rbrace$, $2=\\lbrace 0,1\\rbrace$, and $3=\\lbrace 0,1,2\\rbrace$. Each natural number is the set of all smaller natural numbers. The empty set therefore does not “produce something from nothing” in a metaphysical sense. Mathematics begins with a formally defined object that has no elements, then uses axioms and operations to generate further structures. Numbers are not rearrangements of the empty set; they are relational constructions built from the empty set within a formal system.\nThe same point extends to the rest of mathematics. Integers can be constructed from pairs of natural numbers, rational numbers from pairs of integers, real numbers from rational numbers, and calculus from real numbers, functions, limits, derivatives, and integrals. On that basis, calculus can in principle be reduced to set-theoretic constructions. But the empty set alone is not enough. The foundation of standard set theory is not simply $\\varnothing$, but the logical framework, the primitive membership relation $\\in$, and the axioms that govern membership. The empty set is not the deepest irreducible primitive; it is the first distinguished object whose existence is guaranteed by the axioms.\nFrom mathematics, the question moves to the natural universe. The universe may be understood as fully determinate physical reality. Perfection, in the strongest sense, would be the totality of all truths about that reality. A theory is an axiomatizable formal representation, while a model is a mathematical structure satisfying such a theory. A set-theoretic model may preserve the structure of the universe without making the universe literally a set. A map can preserve the structure of a city without being made of streets and buildings.\nStructural realism provides the philosophical bridge. Under an ontic structuralist view, the universe is not fundamentally a collection of objects that later enter into relations. Rather, the universe is the network of relations itself, while “objects” are stable positions or placeholders within that relational structure. An electron, for example, is not an independent substance hidden behind its properties. It is identified by its role in a web of relations: charge, spin, mass, lepton number, coupling strengths, symmetry behavior, and interactions with fields. If all these relations were removed, physics would have no remaining criterion by which to identify the electron. The electron’s identity is exhausted by the role it plays in the whole structure.\nA hypothetical lepton-antilepton universe helps test the limits of the relational view. A universe containing exactly one lepton and one antilepton may have globally balanced quantum numbers, such as zero total electric charge or zero total lepton number. But global cancellation does not mean absence of structure. The particles exist as distinct physical states because of their relational asymmetry: their positions, states, and behavior relative to one another within the total physical structure. Localization and distinguishability are not isolated essences hidden inside the particles themselves. They are features of the relational network in which the particles appear. Objects exist relationally within the universe; the universe exists as the total relational structure.\nThe relational picture raises the question of what remains invariant across different successful representations. A set-theoretic model uses membership as its primitive relation, but a category-theoretic or type-theoretic model may use different primitives. If all these formalisms describe the same physical universe, then the true content cannot be the specific symbol $\\in$ or the empty set itself. The better candidate for reality is whatever remains unchanged across all adequate representations: symmetries, conservation laws, causal structures, dynamical relations, and transformation behaviors. Such features are not merely labels used by one formalism. They are structural relationships that any successful representation must encode.\nGödel’s incompleteness theorem introduces a limit when such invariant structure is captured in a formal theory. If a mathematical theory of the universe is consistent, recursively axiomatizable, and powerful enough to express basic arithmetic, then it cannot prove every truth expressible in its own language. That limitation does not mean physical reality itself is unstable, incomplete, or broken. The universe may remain fully determinate. The limitation belongs to the formal description, not necessarily to reality. An axiomatizable model cannot be identical to perfection if perfection means the totality of all truths. Truth in the universe and provability within a specific formal system remain distinct.\nThe gap between syntax and truth matters most where reality is not merely described, but experienced. Suppose the relational structure of the universe reaches a threshold of complexity where subjective, first-person experience emerges. A physical theory may chart every relation, every state transition, and every functional interaction within the system. Even then, such a theory may fail to capture the qualitative fact of what it is like to be that structure. Gödel’s theorem does not prove that consciousness is unprovable. It only shows that truth and proof can come apart in any sufficiently powerful formal system. Consciousness may be one of the places where the difference between formal description and lived reality becomes impossible to ignore.\nA structural variant of the hard problem follows from that possibility. The difficulty may not be merely that current science lacks the right instruments or enough information. The difficulty may be that a formal representation of physical relations does not automatically yield the first-person reality of experience. From the first-person point of view, experience is given before theory begins. A model may describe the physical conditions under which consciousness appears, but describing those conditions may not be the same as deriving the subjective truth of experience from formal primitives. The hard problem may therefore mark a boundary between structural representation and qualitative presence.\nThe entire argument turns on a layered distinction. Mathematical structures can be reduced to formal primitives, but those primitives are not ordinary objects inside the theory. Physical objects can be treated as relational positions, but the universe as a whole is not an object inside a larger container. A complete reality may be fully determinate, but any consistent, arithmetic-capable axiomatizable model of that reality will leave some truths unreached by proof. Consciousness may be true even if a formal model cannot exhaust what that truth means from within its own syntax. The resulting view is neither simple formalism nor simple realism. It is a structural view constrained by incompleteness: the universe may be a pattern of relations, but no formal representation of that pattern can automatically claim to be its final perfection.\n","permalink":"https://jankmetko.fyi/personal/the-relational-universe/","summary":"Can reality, including consciousness, be fully captured by formal structure?","title":"The Relational Universe: Structure, Incompleteness, and the Limits of Formalism"},{"content":"Artificial intelligence has become good at doing things people would rather not do. It reads manuals, checks mathematical proofs, summarizes books, and writes software. Every year, another task once requiring effort becomes something we delegate to a machine. Such progress raises a philosophical question: if intelligence can increasingly be outsourced, what remains uniquely ours?\nOne answer is that AI may force human inquiry inward. As machines take over more external tasks, the deepest remaining question may become not what we can do, but what we are. Yet that question may contain a built-in limit. Consciousness may never fully explain itself because the conscious system doing the explaining is also the system being explained.\nAt first, the answer to what remains ours seems obvious: our goals. AI performs tasks while we decide which tasks matter. As machines assume more practical work, the question shifts from how we accomplish our goals to why we pursue them. Perhaps the only thing we cannot outsource is deciding what ultimately matters.\nExperience appears to provide the first answer. We travel to visit someone we love, work to enjoy freedom, and study to satisfy curiosity. Yet people rarely settle on a fixed set of experiences. Psychologists describe hedonic adaptation as our tendency to normalize the extraordinary: the child who dreams of a bicycle eventually finds riding ordinary, and the scientist who solves one problem searches for another. Human life therefore resembles an expansion of what counts as meaningful experience rather than a search for permanent pleasure.\nThat expansion has often moved outward, enlarging our concern from ourselves to family, community, humanity, and perhaps the universe. Yet AI may now be pushing us to look inward. As machines inherit more of our external cognitive tasks, consciousness is turning back toward the process that makes thought possible.\nThat inward turn presents an unusual scientific problem. Astronomers study galaxies, but galaxies do not participate in the investigation. Chemists study molecules, but molecules do not attempt to understand chemistry. Consciousness is different. The system performing the explanation is also the system being explained.\nMany discussions invoke self-reference at this point. Gödel\u0026rsquo;s incompleteness theorems, self-referential paradoxes, and the image of a map containing itself are often cited as reasons why complete self-understanding might be impossible. Those analogies are useful, but they do not settle the issue. Science rarely explains phenomena by reproducing every detail. Meteorologists explain hurricanes without simulating every molecule of a storm, so explaining consciousness should not require the brain to compute every neural interaction occurring inside itself in real time.\nRejecting the need for perfect simulation does not eliminate the puzzle. The real question becomes: why might a conscious system still fail to fully explain itself?\nOne possible answer lies in how self-modeling works. Brains operate under limits of energy, memory, and time. Evolution favors systems that regulate themselves efficiently rather than perfectly. A brain attempting to monitor every microscopic event within itself would spend its resources observing instead of acting. Efficient regulation therefore requires abstraction.\nThe resulting abstraction resembles a computer interface. A desktop icon is not a tiny object hidden inside the machine; it is a simplified representation that lets users interact with complex machinery. Likewise, the colors we perceive, the emotions we feel, and the sense of being the same person through time may function as interfaces rather than exhaustive descriptions of brain activity.\nAn interface is designed for function, not completeness. A road map omits the chemistry of asphalt because that information is irrelevant to navigation. Similarly, a conscious self-model may omit the details of neural computation because they contribute little to survival or decision-making. The omission may be what makes rapid cognition possible.\nA natural objection follows. If consciousness relies on simplified internal models, why can\u0026rsquo;t neuroscience uncover the hidden mechanisms? The objection separates two questions. One asks whether science can explain the physical processes producing consciousness. There is no obvious reason to think such an explanation is impossible. The other asks whether a conscious system can completely represent its own first-person perspective while remaining that perspective.\nImagine a neuroscientist studying another person\u0026rsquo;s brain. The scientist may understand every neural mechanism involved in conscious experience, yet the explanation remains objective. The conscious subject possesses direct first-person access but cannot simultaneously inspect the neural computations producing that experience. For the subject to peer beneath the desktop icon and inspect the raw code of their own neurons would defeat the purpose of the abstraction. One perspective offers explanation without experience; the other offers experience without complete explanatory access.\nThe difference does not imply that consciousness is supernatural or beyond science. The difference may instead reveal a limit faced by finite systems that model themselves. A system that regulates itself through an internal model cannot replace that model with the complete process being modeled without losing the advantages the model provides. The representation succeeds because it is not identical to the mechanism.\nThe proposal is therefore modest. The proposal does not claim that consciousness is nonphysical, nor does it solve the Hard Problem by explaining why experience exists. Instead, the proposal asks whether any finite system capable of subjective experience may face limits in representing its own first-person character from within. The obstacle, if it exists, comes from a simple conflict: the system is trying to understand itself while also being the thing it is trying to understand.\nThe oldest philosophical question—\u0026ldquo;Who am I?\u0026quot;—therefore acquires a different interpretation. Perhaps the answer is not merely unknown because science has not advanced far enough. Perhaps complete self-understanding requires something no finite self-model can achieve: serving as the observer, the observed, and the complete representation of both at the same level of description.\nSuch a conclusion is neither pessimistic nor anti-scientific. Every advance in neuroscience, psychology, artificial intelligence, and philosophy can deepen our understanding of consciousness. Yet even a complete objective science of the mind may leave intact the local perspective from which consciousness is experienced.\nPerhaps that remaining perspective is not evidence that consciousness lies beyond science. Perhaps it marks the boundary every self-modeling system eventually reaches. If so, consciousness is unique not because it presents an unsolvable mystery, but because it is the only part of the universe we know that persistently attempts to understand itself from the inside.\n","permalink":"https://jankmetko.fyi/personal/can-i-explain-myself/","summary":"AI may leave us with only ourselves.","title":"Can Consciousness Ever Explain Itself?"},{"content":"Free will usually seems obvious: I decide, and then I act. But if consciousness depends on the brain, what would it take for a choice to be genuinely free rather than the result of prior physical events?\nLet us formulate a possible answer called Agent-Causal Interactionist Dualism, or ACID, and then ask whether the model preserves free will.\nRaise your hand.\nIt seems like one of the simplest things in the world. Before your hand moved, you decided to raise it. Then it rose. The sequence seems obvious: first came the thought, then the action.\nSuppose someone watched you. They could not directly observe your thoughts, but they would conclude that you intended to move before doing so. Your hand did not leap upward by itself; something preceded the movement. We naturally assume a simple chain of events: conscious intention leads to physical action.\nBut what exactly is that intention?\nModern neuroscience reveals a close connection between consciousness and the brain. Everything we know suggests that human consciousness requires a physical substrate, and that the substrate is the neural architecture of the brain.\nSo, tying consciousness to the physical brain seems to solve the mystery. Perhaps the thought of raising your hand is another physical process occurring in a neural network.\nThe problem appears when we adopt a stronger, libertarian notion of free will. Suppose free will means that your decision is not the inevitable consequence of previous physical events. Suppose, given exactly the same circumstances, you genuinely could have chosen otherwise.\nIf every event in your brain is determined by prior physical states and the laws of nature, then your decision to raise your hand was fixed before you became aware of it. Your feeling of freedom would itself be another consequence of physical laws.\nIf we want to preserve that stronger notion of freedom, the physical substrate alone seems insufficient. Suppose consciousness is not reducible to the brain.\nSuppose there exists a non-physical conscious self—a soul—that depends on the brain as its interface with the material world but is not identical to it. If such an agent could initiate actions not fully determined by prior physical states, we would have the beginnings of a model capable of preserving strong free will.\nLet us combine a conscious agent, causal interaction between mind and matter, and a separation between the two into a framework that we will call Agent-Causal Interactionist Dualism, or ACID.\nThe ACID model rests on the following assumptions:\nConsciousness is a non-physical soul. The soul requires a physical substrate, the brain, to exist or operate. The physical substrate obeys physical laws. The laws of physics are not fully deterministic. The physical world is not causally closed; non-physical causes can influence physical processes through the brain. The soul can initiate voluntary actions through the physical substrate. Once initiated, bodily actions unfold according to physical laws. The soul is an irreducible agent capable of originating actions that are neither determined by prior states nor random. Voluntary actions originate in the soul. Causes precede their effects; therefore, conscious intentions precede voluntary actions. At first, the framework seems to preserve everything we want. Conscious experience remains real, the brain remains essential, and strong free will survives.\nBut now the harder question appears:\nWhere does the soul first affect the brain?\nConsider the thought:\nI will raise my hand.\nIts meaning does not reside in a single neuron. Meaning is relational. Meaning depends on connections among pieces of information distributed through a neural network.\nNeurons are physically separated, and signals require time to travel between them. A complex representation cannot appear fully formed everywhere at once.\nOne possibility is that the soul influences many distant neurons simultaneously. Another is that the soul affects a small region of the brain and allows neural processes to spread the consequences through the network.\nThe second possibility is simpler because it requires fewer assumptions and lets the brain\u0026rsquo;s connectivity do most of the work.\nAwake brain surgery offers a useful analogy, though not evidence for the soul. Neurosurgeons can stimulate tiny regions of the brain with electrodes while patients remain conscious. Patients may report a memory, an emotion, or an urge to move. A physicalist could say that the brain can generate such experiences mechanically.\nStill, the example reveals something important about the brain. Small perturbations can produce large, distributed effects.\nIf ACID is correct, the soul would probably have to exploit a similar amplification structure. The soul would not need to inject an entire thought into the brain. It would need only to trigger or bias existing neural dynamics.\nThe preceding observations make the localized-interaction version of ACID more economical. Instead of imagining the soul writing a complete message across the brain all at once, we can imagine the soul initiating a small change in a sensitive region of the neural network. The brain would then amplify and distribute that change through physical processes.\nBut that answer introduces another difficulty.\nIf the soul is non-physical, where is it?\nDoes it occupy space? If it occupies space, it begins to resemble an unusual kind of matter. If it does not occupy space, then how does it interact with a physical brain at all?\nTouching, reaching, and influencing all assume spatial relationships. A non-spatial soul cannot sit next to a neuron or push an ion through a membrane.\nAt this point, ACID reaches its explanatory limits.\nThe model can say that the soul stands in some primitive relationship with the brain, where changes in the soul correspond to changes in neural tissue. But the model cannot explain how that interaction occurs.\nACID preserves free will by placing the origin of voluntary action outside the physical chain, but it does so by introducing a relation we do not understand.\nACID also requires us to give up the idea that physics alone provides a complete causal account of everything that happens in the physical world. The laws of physics may still govern matter once physical processes are underway, but they would not be the whole story. Conscious choice would have to enter the physical world through a special interface between soul and brain.\nNone of these problems prove that ACID is wrong. Nature is under no obligation to be elegant. In fact, ACID may preserve the kind of free will many people believe they possess. The model allows us to say that when we raise our hand voluntarily, the conscious intention comes first and the action could have been otherwise.\nYet each attempt to protect that intuition introduces another assumption and another mystery. We begin with a familiar experience:\nI decide, and then I act.\nTo preserve that experience in its strongest form, we introduce a non-physical soul, a physical brain substrate, cross-substance causation, agent origination, an incomplete physical causal order, and a primitive interface between the non-spatial and the spatial.\nPerhaps the most difficult question is not whether a theory of free will can be made internally consistent. ACID is.\nThe more difficult question is whether we are willing to pay an increasingly high theoretical price to preserve the idea that, when we raise our hand, we truly could have done otherwise.\nAnd we are not.\nSo ACID is probably not the way.\n","permalink":"https://jankmetko.fyi/personal/free-will-strange-assumptions/","summary":"Explaining free will may require assumptions too strange to accept.","title":"If Free Will Exists, What Would It Have to Look Like?"},{"content":"Human beings fear their own death. Other species show survival behaviors, but human beings can go further: they can picture a future in which they irreversibly no longer exist. Understanding the consequences of death creates a dread that instinct alone does not explain. The fear is not of the process of dying, but of the disappearance of the experiences. The question then becomes deeper than biology: why should a conscious system value its own continued existence, and what exactly is threatened when death approaches?\nDoes survival answer that question? No. Evolution explains why organisms that preserve themselves are more likely to reproduce, and many human goals support that process. Hunger preserves the body, curiosity gathers knowledge, social bonds strengthen cooperation, and reproduction extends genetic lineage. Ok, survival is treated as an end, but for what purpose? If every goal serves self-preservation, and self-preservation serves only itself, then conscious life appears trapped in a self-referential loop. Consciousness not only tries to survive but also models itself, seeking to understand both the external world and its own existence.\nOther systems can be self-referential and have no purpose. Cream swirling through coffee forms eddies that persist before dissolving. Computer programs can model parts of their own execution without intention. These examples show that recursion alone does not create meaning. Consciousness differs because representation is accompanied by experience. A human being not only models the self but inhabits that model from the inside. The unanswered question is therefore not simply how a system represents itself, but why representation is accompanied by the fact that there is something it feels like to be that system. Is the ability to be “self-aware” responsible for the “fear of death”?\nThis point raises the possibility that solving the hard problem of consciousness would also clarify other philosophical puzzles. Questions such as \u0026ldquo;Why do I fear death?\u0026rdquo;, \u0026ldquo;Who am I?\u0026rdquo;, \u0026ldquo;What does it feel like to be me?\u0026rdquo;, \u0026ldquo;Why does time seem to flow?\u0026rdquo;, and \u0026ldquo;What is my purpose?\u0026rdquo; all require a first-person perspective. They seem tied to the same center. If consciousness explains subjectivity, then perhaps these questions are different ways of examining the same phenomenon. Another possibility is more troubling: consciousness may explain why these questions arise while still unable to answer them, because every attempt at explanation begins inside the very system being examined.\nIf this limit is real, the deepest questions of mind, including why we fear death, may resemble Gödel\u0026rsquo;s incompleteness phenomenon in a loose philosophical sense. The comparison is not mathematical. It is structural. Just as formal systems can run into truths that cannot be proved from within their own rules, consciousness may run into questions about itself that cannot be settled from within its own representational structure. Such a conclusion is unsatisfying, not because it disproves purpose, but because it suggests that purpose may remain unreachable from the inside. Faced with this limit, the idea of an absentee creator becomes attractive. Such a creator need not explain or intervene; imagining that someone understands the whole picture is comforting. But the idea of god moves the mystery rather than solves it. The question of purpose shifts from humanity to the creator, and the regress remains.\nThe existential difficulty becomes sharper from the first-person perspective. Every conscious mind occupies a private world of experience inaccessible to any other subject. From that viewpoint, the statement \u0026ldquo;When I die, my universe ends\u0026rdquo; is almost tautological. The physical universe may continue, but every star, friend, equation, and memory has only ever appeared through the window of one\u0026rsquo;s own awareness. When I cease, no further appearances remain. Death represents not only the loss of future experiences, but the disappearance of the only reality I was ever given. From this internal horizon, it is hard to distinguish the world continuing after one\u0026rsquo;s death from the world ending altogether. Legacies, future generations, and contributions seem to lose their foundation if there is no remaining subject for whom those achievements matter. However, this isolation reveals a paradox: the internal model that makes up \u0026ldquo;my\u0026rdquo; universe was never built in isolation. The language we think with, the concepts we use to fear death, and the mathematics we use to map reality are inherited. We are individual nodes in a network of consciousness that came before us and will outlast us. While our private world of experience ends at death, our minds are built inside a shared structure of understanding.\nHuman civilization shows that understanding is a collective achievement. No individual could build a modern computer, discover the Standard Model, or develop contemporary mathematics from scratch. Knowledge emerges from many connected minds whose shared efforts exceed any individual mind. More importantly, discovery is rewarding in itself. If consciousness is a representational system, then survival is not its deepest goal but the condition that makes continued representation possible. A living system preserves itself because only a continuing subject can keep enlarging its understanding of reality. Other conscious minds matter not merely as partners in survival, but because they are independent centers of experience—other windows looking onto the same cosmos, each revealing aspects of reality inaccessible to any single observer.\nThe argument converges on a single insight: understanding and imagination are the two prominent activities of consciousness. Understanding enlarges the internal model of reality, while imagination expands the space of realities that can be represented. Science, philosophy, art, relationships, and civilization are expressions of these two activities, each extending what consciousness can know and envision. If consciousness is a system that represents itself, then its highest expression may not be the possession of final answers, but the refinement of its understanding. Perhaps consciousness exists not because every mystery can be solved, but because there is a world that can be understood at all.\n","permalink":"https://jankmetko.fyi/personal/fear-of-death/","summary":"What if death is not the end of the universe, but the end of the only universe you ever knew?","title":"When My Universe Ends"},{"content":"At first glance, the universe appears to move in a clear direction. We begin with simple constituents—particles, atoms, and molecules—and over time we see the emergence of stars, planets, chemistry, and eventually life. From single cells come complex organisms, then animals, and eventually humans, who build language, technology, and civilizations capable of reshaping entire planets. It is natural to summarize this history as an upward trend in which matter becomes progressively more organized and complex.\nPhysics seems to partially support that intuition through the second law of thermodynamics, which states that total entropy increases in closed systems. Yet within that overall trend, we observe local regions where structure and order increase dramatically. Living cells, ecosystems, cities, and machines appear as islands of organization within a broader tendency toward disorder. That observation can suggest that the universe is locally “fighting” entropy to produce structure.\nThe key correction begins with thermodynamics. The second law applies to closed systems, stating that total entropy increases overall. However, it does not forbid local decreases in entropy. In fact, local increases in order are not exceptions but expected features of systems through which energy flows. The important point is that local structures are always paid for by larger increases in entropy elsewhere.\nFor example, sunlight continuously pours energy onto Earth, creating strong energy gradients between the Sun, the Earth, and space. Those gradients allow weather systems, oceans, and life itself to form. These systems develop internal structure while simultaneously producing waste heat that is radiated into space. The key point is not that structure resists entropy, but that structured configurations arise as intermediate pathways through which energy flows from high gradients to equilibrium. They are not fighting entropy; they are part of the mechanism by which entropy increases globally.\nThat principle is captured in physics by the concept of dissipative structures. Hurricanes, convection cells in boiling water, crystals forming as liquids cool, and living organisms are all examples. None of these systems are striving for order; they persist because they are stable configurations under continuous energy flow.\nBiological evolution introduces a similar but more subtle illusion of direction. It is often described as a progression from simple organisms to complex ones, but it has no goal or preferred direction. It consists only of variation, inheritance, and selection, where successful genetic configurations persist and others disappear. Complexity can arise, but only when it is locally useful for survival in a given environment. There is no inherent drive toward intelligence or sophistication; instead, there is a filtering process in which only configurations that remain viable under environmental constraints persist over time.\nHuman civilization extends the same pattern into a new domain. Language, science, institutions, and technology spread through cultural transmission, allowing information to replicate and evolve far more rapidly than genetic evolution. However, cultural evolution does not introduce a new independent force of selection. Cultural evolution still depends entirely on human cognition, behavior, and physical energy flows. It is best understood as biological evolution extending its information processing outside the genome, not as a separate evolutionary engine.\nAcross all three domains—physics, biology, and civilization—the same principle appears: systems persist only if they are stable under their environment. Unstable configurations vanish quickly, while stable ones accumulate history and become disproportionately visible. This selection effect means that we only observe the patterns that survive long enough to be observed. Complexity appears to increase over time because simple configurations are often either too stable to notice change or too fragile to persist in evolving conditions, while a narrow set of complex configurations happen to be both stable and capable of maintaining energy flow.\nOnce this selection effect is recognized, the central correction follows: the universe does not evolve toward complexity, optimize anything, or progress toward any goal. It consists of local interactions governed by physical laws, continuous energy flows, and constant formation and destruction of structure, with only a small subset of configurations persisting under those constraints. What appears as progress is simply the survival of structures that remain viable under environmental filtering.\nEven after accepting this conclusion, the sense of direction remains compelling because observers are part of the same system. Human brains are pattern-compressing, narrative-building mechanisms that infer intention from regularity and convert long chains of survival into coherent stories. As a result, we naturally interpret stable and self-sustaining structures as purposeful or directed, even when no such direction exists in the underlying physics.\nThe result is a final inversion of the initial intuition: the universe does not organize itself. It only contains temporary patterns that persist long enough, and in sufficiently structured ways, to make organization appear as if it were a fundamental tendency.\n","permalink":"https://jankmetko.fyi/personal/the-not-directed-universe/","summary":"It feels like everything is moving toward something.","title":"Why Does the Universe Look Like It Is Directed?"},{"content":"From cells, to monkeys, to you, to what’s next?\nPeople first learn about evolution in biology class: living organisms change over time through variation, selection, and heredity. Mutations create differences, environments “select” which traits survive, and successful traits are passed on. Over long periods, evolution produces everything from bacteria to humans.\nBut biological evolution is only one instance of a broader pattern. Once you start looking closely, variation, selection, and heredity appear in many systems that are not biological. In culture, ideas vary as people generate new thoughts, some spread widely while others disappear, and knowledge is preserved through language and writing. In technology, different designs are created, some are adopted because they are useful, and successful designs are copied and improved. In science, competing theories are tested against evidence, with better explanations surviving and weaker ones fading out. Across these cases, evolution is not only about life; it is about any system that produces copies with variation under selection pressure.\nFollowing the pattern further reveals a larger historical sequence. Chemistry contains primitive forms of selection among replicating molecules. From chemistry, biological evolution emerges through genes. Nervous systems then introduce learning within a lifetime. With humans, language allows ideas to be transmitted independently of genes, accelerating the accumulation of information. Civilization extends language through writing, science, and institutions. Today, computers and artificial intelligence extend the pattern further, processing and generating information at greater speeds than biological systems.\nAt first glance, evolution might seem tied to biology, as if genes are the main drivers and organisms are their vehicles. But with human intelligence, a shift occurs: systems begin to operate on faster timescales than genetic change. Scientific reasoning, engineering, and AI replace blind variation with guided search. Scientists build models rather than guess randomly, and engineers simulate rather than rely on trial and error. Selection is also partially internalized, since many possibilities can be evaluated in advance through prediction rather than physical survival alone. The shift turns evolution from a blind process into a more directed form of optimization over possibilities.\nAs cultural and technological systems grow more capable, the shift intensifies. Artificial intelligence may begin to design improved versions of itself and generate new knowledge at speed, reducing dependence on human cognitive limits. The same underlying structure of variation, selection, and heredity remains, but it increasingly operates in fast, self-referential digital systems rather than slow biological ones.\nThe argument leads to a final question: what happens when a general process of information optimization becomes able to understand and modify its own rules? At that point, we are no longer observing evolution from outside. We are living inside it as evolution begins to reshape itself.\n","permalink":"https://jankmetko.fyi/personal/the-selfish-gene/","summary":"From cells, to monkeys, to you, to what’s next?","title":"When Evolution Leaves Biology Behind"},{"content":" My name is Jan Kmetko. I teach at Santa Rosa Junior College in California, and in the summers I volunteer at Mount Wilson Observatory. I love physics and astronomy.\nI started this blog to express my ideas about the universe and life without making the posts needlessly academic.\nThe site has two main areas:\nTeaching — learning tips and selected lessons. Personal — essays and reflections on philosophy, science, astronomy, and life. I write for curious readers who enjoy careful thinking in an informal format. My personal essays are classified by reading depth: choose light articles for casual reading and deep articles for heavier mental engagement.\nIf you would like to make a public comment or ask a question about one of my posts, email me. I may include your comment under the essay, either anonymously or with your name.\nStart reading!\n","permalink":"https://jankmetko.fyi/about/","summary":"\u003cp\u003e\u003cfigure\u003e\n    \u003cimg loading=\"lazy\" src=\"mugshot.png\"\n         alt=\"Photo of Jan\" width=\"280\" height=\"280\"/\u003e \n\u003c/figure\u003e\n\n \u003c/p\u003e\n\u003cp\u003eMy name is Jan Kmetko. I teach at \u003ca href=\"https://chemistry-physics.santarosa.edu/staff/jan-kmetko\"\u003eSanta Rosa Junior College\u003c/a\u003e in California, and in the summers I volunteer at \u003ca href=\"https://www.mtwilson.edu/soar/\"\u003eMount Wilson Observatory\u003c/a\u003e. I love physics and astronomy.\u003c/p\u003e\n\u003cp\u003eI started this blog to express my ideas about the universe and life without making the posts needlessly academic.\u003c/p\u003e\n\u003cp\u003eThe site has two main areas:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href=\"/teaching/\"\u003e\u003cstrong\u003eTeaching\u003c/strong\u003e\u003c/a\u003e — learning tips and selected lessons.\u003c/li\u003e\n\u003cli\u003e\u003ca href=\"/personal/\"\u003e\u003cstrong\u003ePersonal\u003c/strong\u003e\u003c/a\u003e — essays and reflections on philosophy, science, astronomy, and life.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e \u003c/p\u003e","title":"About Jan Kmetko"}]