Posts by paulralph@mastodon.acm.org
 (DIR) Post #ASTX56gwfSA6rzx5w8 by paulralph@mastodon.acm.org
       2023-02-08T16:37:37Z
       
       0 likes, 0 repeats
       
       @brewsterkahle SCI-HUB is so important to humanity's continued survival that its ostensible illegality casts doubt upon the validity of law itself.
       
 (DIR) Post #AWcXDUAfWsHT9qqYXw by paulralph@mastodon.acm.org
       2023-06-12T14:14:14Z
       
       0 likes, 0 repeats
       
       So... the #apple #visionpro scans you to build a detailed 3D model of your face, so it can generate a convincing deepfake avatar *of you* for FaceTime. And where, exactly, is this model stored? What can Apple do with it? What could other bad actors do with it if it were stolen?
       
 (DIR) Post #AWiSN5quIOkZhsy8Om by paulralph@mastodon.acm.org
       2023-06-14T11:40:42Z
       
       0 likes, 0 repeats
       
       @seresearchers today at #EASE2023 I learned that CPU time is not always a good proxy for energy consumption in mobile apps because mobiles have both efficient and performance cores, so it depends which cores are being used.
       
 (DIR) Post #AWiSN6YrexubuDXEum by paulralph@mastodon.acm.org
       2023-06-15T07:00:33Z
       
       0 likes, 1 repeats
       
       @seresearchers The plot thickens! It seems that for any machine with a multi-core processor, CPU time is not a good proxy for energy usage because the latter depends on how many cores are used. This has important implications for research on software sustainability, in which energy usage is a key dependent variable.
       
 (DIR) Post #AWiSzVMnsk0rqJCjBI by paulralph@mastodon.acm.org
       2023-06-15T13:02:07Z
       
       0 likes, 0 repeats
       
       @mtorchiano @seresearchers There's a thesis in that :-)
       
 (DIR) Post #AY80YpFIMhSSNQfBxI by paulralph@mastodon.acm.org
       2023-07-27T13:49:42Z
       
       0 likes, 1 repeats
       
       @seresearchers @academicchatter What do you use for stats? Is writing data analysis scripts manually (e.g. in R or Python) really better than using a stats program like SPSS? Has anyone actually studied relative error rates or is it all just personal opinion? I just asked chatGPT to write a python script to perform an ANCOVA and its output was totally wrong. I bet novice and probably intermediate analysts make more mistakes with Python than SPSS.
       
 (DIR) Post #AZ3jAtvzjrjnIR0LZ2 by paulralph@mastodon.acm.org
       2023-06-29T14:32:39Z
       
       1 likes, 0 repeats
       
       @SIGSOFT This seems tentatively good, but reading about PACM raises many questions. If FSE is moving to two-phase review, what's the plan to protect authors from reviewers demanding unreasonable changes? Does this mean authors no longer have to attend the conference in person? Will FSE stop price gouging virtual author-attendees? To what extent does PACM control FSE's review process now?
       
 (DIR) Post #AZoim4qOcArNfrTfzE by paulralph@mastodon.acm.org
       2023-09-15T20:10:19Z
       
       0 likes, 1 repeats
       
       Journals: Why won't anyone accept our #peerreview  invitations?!?!Also journals: please spend 30 mins updating your reviewer profile, research interests, conflicts, password, contact details and availability before writing your review. Oh, and check out our terms of service! You'll love this bit about our right to use your review text to train AI models. @academicchatter