From: kymhorsell@gmail.com   
      
   EXECUTIVE SUMMARY   
   - Space telescope images seem to be edited to remove "certain things"   
     from the view of scientists in the US and around the world.   
   - The editing follows 2 patterns -- the day before a UFO flap is   
     registered at e.g. NUFORC the number of images acked in public   
     databases is consistently smaller than usual. We suspect some   
     images have been removed.  For those images remaining, in areas of   
     the sky and at times were "certain things" can be predicted to be   
     likely seen, images seem to be consistently blacked out. Leaving the   
     empty file in the database may be a method to avoid holes in the   
     database being observable to casual view. Only after uploading the   
     full-sized (in this case 30+ MB) file does the researcher find it is   
     blacked out apart from the information in the file header that   
     otherwise gives the meta info for the data that was once there.   
   - We find there is a statistically strong link between the number of   
     images listed in log files for at least one telescope and reported   
     UFO activity the following day. Images taken near the position of   
     Neptune (RA near 0 or 360 for most of the period the telescope has   
     been operating) on key dates also seem to be consistently "blacked out".   
   - The stats s/w finds the best match-up between UFO counts and   
     image counts shifts the UFO data back 1 day and takes logs.  A log   
     model is consistent with the problem of N objects randomly allocated   
     to M images. The fraction of images without any of the objects in it   
     would tend to a function of log(N).  So both the time-shift in the   
     model and the functional form is consistent with someone trying to   
     hide something related to contemporaneous UFO activity.   
   - We can't assign blame to NASA or govt scientists. It is perfectly   
     possible they all act in good faith and do not notice these patterns   
     in their everyday work. NASA administrators must be aware they   
     operate in an informal branch of the military and are directly   
     subject to national security concerns. They may even know that   
     certain steps are taken as a matter of course without being fully   
     aware what those steps might entail.  But the effect of "invisible   
     changes" to public data may threaten to invalidate a growing amount   
     of research that relies on it.  While scientists reaching certain   
     wrong conclusions is likely "part of the plan", the errors are not   
     contained to just that part of scientific work. No matter how   
     careful and planned the editing, an obvious and spreading bias has   
     been introduced. If this same methodology is used in all govt   
     science (for the same or even other reasons) that research can't be   
     regarded as totally reliable and the gobs of public money that go   
     into it are intentionally "partly wasted".   
      
      
   I mentioned some time back I'd found some evidence that images   
   available in key NASA databases seem to have been tampered with to   
   remove evidence of "certain things".   
      
   While NASA administrators maintain the organisation is a straight die   
   it may well be the case that some actions are undertaken with or   
   without its cooperation to remove "material of national security   
   significance" from NASA's public data. More interestingly, some of the   
   public involved are other science organisations around the world.   
   Whoever is doing what is also tampering with scientific research.   
   Removing evidence with "security implications" inevitably is biasing   
   the types of data scientists everywhere get to use to advance their   
   understanding of the universe. That understanding therefore becomes   
   more and more skewed from reality over time. The tampering is   
   producing arguable damage to society.   
      
   But before I fly off the handle and REALLY start frothing at the mouth, let's   
   look at the evidence. See if you find it as convincing as I do.   
      
   In researching the goings-on in the sky over N America and my own neck   
   of the woods I began looking at telescope data gathered by several   
   organisations and available via the web. Initially I looked at light   
   curves from individual stars. That work found that the average light   
   from stars in certain parts of the sky varied suspiciously like the   
   activity reported much closer to the ground in the sky over N America.   
   The patterns were so clear you could work up a predictive model that   
   would tell you to quite high certainty when and where such near-ground   
   activity would occur just by looking at the avg brightness of the   
   stars in certain parts of the sky over the course of hours before that   
   activity was due to happen.   
      
   Using a simple machine learning model my programs predicted from the   
   initial upoloadings of 1000s of light curves where and when and in   
   what direction you should be able to look to find even more   
   interesting things than just the odd bunch of stars suddenly going a   
   little dimmer over the course of a few hours and then going brighter   
   again.  At that time -- and this is the way AI stuff works -- I had no   
   real simple mental model of what the machine learning algorithm had   
   found.  It just told me get data about this part of the sky at this   
   time and you will see something interesting.   
      
   So I expanded my uploads to include actual full-frame images of   
   parts of the sky gathered by various space telescopes. The most   
   convenient for my monthly Internet bill were TESS images that came in   
   30 MB FITS files (FITS is a format beloved of astronomers to bundle up   
   all the information about some part of the sky and can include tables,   
   gobs of meta-information about the camera and telescope used, and even   
   tables of sub-images and graphs as well as the meat-in-the-sandwich   
   32-bit gray-scale detailed images of the part of the sky in question;   
   it's like the astronomers version of a ZIP file).   
      
   So I went about uploading modest numbers of these things from my   
   favourite space telescope archive and storing them on a big disk on my   
   local machine. To allow for possible 1000s or millions of images -- I   
   didn't know how many I would need -- I compressed all the FITS files   
   with a super compression algorithm. That algorithm took those FITS   
   files and reduced them by 50% or so. Pretty good performance given   
   FITS are already somewhat compressed.   
      
   But as the work progressed I suddenly found some of these compressed   
   files were amazingly small. With a normal compressed FITS was 25MB,   
   some of them were only 4KB long! What had happened?   
      
   After some checking it turned out those small FITS files were   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)