[sage-devel] Re: talk
On Tuesday, July 24, 2018 at 4:49:45 PM UTC-5, Timo Kaufmann wrote: > > I really like your wishlist! The all-or-nothing nature of sage and the > slow startup time (although it's actually more like 1.3 seconds with a warm > cache on my machine) are my biggest pain points. > I've encountered incredibly slow Sage startup times on big clusters due to, I believe, the very slow nature of their distributed file systems and the huge number of files Sage opens at startup. This was several years ago, but at the time Sage took several *minutes* to start on the main campus cluster here at UIUC. I ended up using a smaller cluster for my computations where the startup time was "just" 20 seconds. (I was running things from my home directory rather that trying to convince the admins to install Sage on each node which presumably would have fixed things.) >From the viewpoint of SnapPy, a more modular Sage would be fantastic as it would allow us to have more functionality in the stand-alone version and reduce development effort at the same time. Nathan -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: talk
Am Mittwoch, 25. Juli 2018 01:32:38 UTC+2 schrieb William: > > On Tue, Jul 24, 2018 at 5:49 PM, Timo Kaufmann > wrote: > > I really like your wishlist! The all-or-nothing nature of sage and the > slow > > startup time > > (although it's actually more like 1.3 seconds with a warm cache > > on my machine) > > Precisely how are you benchmarking this, and what is your machine? > Can you copy/paste a session? > I haven't done any real benchmarking, just a simple `time sage -c 'quit'`. That takes ~4s on the first run and ~1.3 secs on subsequent runs. My machine is pretty unsectacular: old SSD, old Intel i5 3.2GHz quad. > I'm not sure if its a good idea to separate user and developer error > > messages. > > I think it might be, because most end users of Sage have very, very > different expectations, backgrounds, etc., than Sage developers. > Developers are > often intimately familiar with Python, Cython, and the Sage library. > Instructors who I've talked to who teach undergrads and who don't > give up on Sage typically create their own local "dictionaries" to > help their students figure out what the cryptic Sage error messages > actually mean. I've seen this repeatedly, and it's often done in a > context sensitive way. > > Improving presentation of errors messages (especially the Sage > preparser related mangling) is probably one of the more accessible > projects. Maybe somebody will do it as a GSoC in 2019... > I agree that error messages should be improved. However I think it would be better to improve existing error messages, including information that is useful to users *and* developers. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: talk
On Wed, Jul 25, 2018 at 8:11 AM, Erik Bray wrote: > On Wed, Jul 25, 2018 at 8:03 AM jplab wrote: >> >> Hi, >> >> It is great to have a recent snapshot of the status of Sage and a wishlist! >> >> This summer marks the 10th year when my brother showed me Sage for >> the first time at the beginning of my master (So happy I did not need to >> use Maple or Matematica any more...). It is awesome to see it so >> lively and to be able to participate in improving it! >> >> On the number of contribution graph, it seems that the number increased >> since the move to git in around 2014? Is that accurate? >> >> About moving trac to github, I have been aware of gitlab, an "open-core" >> competitor of gitlab. My institution uses it instead of github as its >> platform >> for internal collaborations. As I am not an expert in this respect, but I >> find >> it interesting to consider, as you mention that your computer was 100% >> open source apart from Magma. Here's a letter from the CEO of gitlab, >> from 2 years ago last week: >> >> https://about.gitlab.com/2016/07/20/gitlab-is-open-core-github-is-closed-source/ >> >> which ends by: >> >> "In conclusion (TLDR), GitLab has an open core business model and ships >> both open and closed source software. GitHub hosts most open source >> projects but ships closed source software." >> >> For what it's worth, I thought I would mention it here... > > We've been keeping quiet about it because we want things to be working > before we start a big discussion / flame war about it but Julian Rüth > and I have already been making some inroads towards enhancing Sage > development on GitLab (as opposed to GitHub), +1 from me for GitLab; it has most of the same important advantages over trac for us. > All this is to say, this is one bullet point in William's slides that > is being worked on. I believe some others are being worked on too, as > able, but progress is slow. I agree with all the main bullet points > (other than the one minor one I pointed out ;) Thanks so much for working on this! -- William (http://wstein.org) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: talk
On Wed, Jul 25, 2018 at 8:03 AM jplab wrote: > > Hi, > > It is great to have a recent snapshot of the status of Sage and a wishlist! > > This summer marks the 10th year when my brother showed me Sage for > the first time at the beginning of my master (So happy I did not need to > use Maple or Matematica any more...). It is awesome to see it so > lively and to be able to participate in improving it! > > On the number of contribution graph, it seems that the number increased > since the move to git in around 2014? Is that accurate? > > About moving trac to github, I have been aware of gitlab, an "open-core" > competitor of gitlab. My institution uses it instead of github as its platform > for internal collaborations. As I am not an expert in this respect, but I find > it interesting to consider, as you mention that your computer was 100% > open source apart from Magma. Here's a letter from the CEO of gitlab, > from 2 years ago last week: > > https://about.gitlab.com/2016/07/20/gitlab-is-open-core-github-is-closed-source/ > > which ends by: > > "In conclusion (TLDR), GitLab has an open core business model and ships > both open and closed source software. GitHub hosts most open source > projects but ships closed source software." > > For what it's worth, I thought I would mention it here... We've been keeping quiet about it because we want things to be working before we start a big discussion / flame war about it but Julian Rüth and I have already been making some inroads towards enhancing Sage development on GitLab (as opposed to GitHub), which we chose in part *because* we believe it to be more politically palatable to the Sage community (this was even before the Microsoft buyout news), as well as the ability to go self-hosted if needed, as well as the nice built-in continuous integration capabilities. I discussed this a bit in a recent ticket: https://trac.sagemath.org/ticket/25914 See also Julian's herculean effort in: https://trac.sagemath.org/ticket/24655 Also, FWIW I should be clear, I personally do *not* want to kill of Sage's Trac any time soon. But I do think there are benefits to accepting contributions through GitLab including but not limited to: 1) More accessible, especially to newcomers, who don't have to learn a new tool / workflow (even if they have not used GitLab before, it's reasonably familiar to anyone who's used GitHub, and allows logging in with their existing GitHub credentials). 2) Just hands-down better code review UX. and eventually, 3) A better continuous integration experience too--the sage patchbot is quite nice, but it will be even nicer to have a *consistent* manner of feedback for changes, whereas sage's informal patchbot fleet is notoriously unreliable.(The buildbot fleet, on the other hand, still has great value in providing slower CI on a wider range of platforms). In order to allow Trac and GitLab to be used simultaneously I've even been working on integration between the two through their respective APIs through Sage's Trac plugin. See e.g.: https://github.com/sagemath/sage_trac_plugin/commit/69aee69c6ac8b0ae67ac483a5085e3e71145bb1c All this is to say, this is one bullet point in William's slides that is being worked on. I believe some others are being worked on too, as able, but progress is slow. I agree with all the main bullet points (other than the one minor one I pointed out ;) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] talk
On Wed, Jul 25, 2018 at 7:54 AM, Erik Bray wrote: > On Tue, Jul 24, 2018 at 9:18 PM William Stein wrote: >> >> Hi, >> >> I just wrote a short talk that I'm about to give at ICMS 2018 about a >> sort of Sage status report and wishlist: >> >> https://goo.gl/qNycb3 [...] > I would just that you also fairly and accurately represent my point, > perhaps by removing that bullet point or at least rewording it. I have removed the bullet point linking to your sage-flame post. -- William -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] talk
On Tue, Jul 24, 2018 at 9:18 PM William Stein wrote: > > Hi, > > I just wrote a short talk that I'm about to give at ICMS 2018 about a > sort of Sage status report and wishlist: > > https://goo.gl/qNycb3 William--I think you do a fine job proving the point of my post on sage-flame by linking to the thread while simultaneously misrepresenting / misunderstanding my point. Perhaps I was unclear and "ponderous" (a perfectly fair assessment!) but where did I write that I dislike the model of voting on topics? All I was saying was that if you're going to respond to a disagreement or open question with a vote it should at least fairly and accurately represent the question being voted on. I would just that you also fairly and accurately represent my point, perhaps by removing that bullet point or at least rewording it. Thanks, E -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: talk
Hi, It is great to have a recent snapshot of the status of Sage and a wishlist! This summer marks the 10th year when my brother showed me Sage for the first time at the beginning of my master (So happy I did not need to use Maple or Matematica any more...). It is awesome to see it so lively and to be able to participate in improving it! On the number of contribution graph, it seems that the number increased since the move to git in around 2014? Is that accurate? About moving trac to github, I have been aware of gitlab, an "open-core" competitor of gitlab. My institution uses it instead of github as its platform for internal collaborations. As I am not an expert in this respect, but I find it interesting to consider, as you mention that your computer was 100% open source apart from Magma. Here's a letter from the CEO of gitlab, from 2 years ago last week: https://about.gitlab.com/2016/07/20/gitlab-is-open-core-github-is-closed-source/ which ends by: "In conclusion (TLDR), GitLab has an open core business model and ships both open and closed source software. GitHub hosts most open source projects but ships closed source software." For what it's worth, I thought I would mention it here... JP Le mardi 24 juillet 2018 21:18:32 UTC+2, William a écrit : > > Hi, > > I just wrote a short talk that I'm about to give at ICMS 2018 about a > sort of Sage status report and wishlist: > > https://goo.gl/qNycb3 > > -- > William (http://wstein.org) > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.