On Wed, Jul 8, 2009 at 12:42 AM, <peterjer...@optushome.com.au> wrote: > On 2009-Jul-06 21:03:20 -0700, William Stein <wst...@gmail.com> wrote: >>> * There is not a support line IT Services can ring up in the event of >>> difficulties installing it on University systems. >> >>But note that there is a mailing list and irc chat room that IT >>Services can get help from. >> >>It might also be worth noting that in the entire history of the Sage >>project, nobody has ever once asked online or to me personally for a >>phone conversation to help them with anything related to Sage. I.e., >>nobody has ever written to sage-support or me personally and said "I >>would like phone support. Is there anybody here who would help?" > > Can I suggest adding a "Commercial Support" link to either the front > page or the existing "Help Groups" page with a "not endorsed by the > Sage Project" disclaimer and allow people and companies to request > they be added to the list. I am thinking of something along the lines > of http://www.freebsd.org/commercial/commercial.html > > This would seem to provide a service for prospective users without > requiring they announce their requirements in a public forum.
I would certainly be OK with such a page for Sage, assuming that there is some discussion before a company is listed there. Some companies aren't very good, and listing them somehow officially endorses them. >>> * Releases happen so frequently that IT depts. cannot hope to keep up >>> with installing the latest releases. >> >>I wonder how much more often Sage releases are than iTunes releases? >>I just checked and our releases are maybe about twice as frequent as >>iTunes. I'm just pointing out that Sage isn't that unusual with its >>release schedule. It used to be 2 years ago though. > > I think it would be more reasonable to compare Sage to the release > schedules for the 3 M's (which I don't know). The 3 M's have the old fashioned traditional closed source release model, which is about once every year or so, with a huge 6-month feature freeze. I much prefer to compare with at least something like iTunes, which is a much more modern software project (in my mind) than the M's. > I _do_ find the current > release schedule overly rapid (though in my case, that's because I'm > trying to do some porting work and I'm finding it difficult to keep > up). Would it be easier for you if you did all your porting work with version of x of Sage, then right when you finished we released version x+1 of Sage and it broke much of what you did? If anything, the regular and rapid release cycle of Sage makes your porting work much much easier than it would be otherwise. What makes your porting work hard is just that Sage *development* (not releases) is rapid. For porting, it is easiest if development of the project your porting is completely stalled, and they are making no progress on new development... :-) For example, this was what Michael Abshoff and I often wanted for Sage during the last 18 months or so, when porting was supposed to be our top priority (that didn't really work out though). We were really against us adding lots of new packages left and right simply because we thought it made porting much harder. I think there have been essentially *no* new external packages added to Sage in a while. The last thing I remember was Spinx and its dependencies, which got voted in last year. Well, Stoll's ratpoints is new, but note that it's a small C program, and an older version of it was already in Sage via mwrank. > It's also worth considering the position of someone who relies > on Sage - every time they upgrade, they need to go through their own > regression test suite to ensure that functionality they rely on has > not been adversely impacted. Again, I would like to argue that this has absolutely nothing whatever to do with the frequency of Sage *releases*. It has everything to do with the rate of Sage *development*. And I think the Sage project has to be developed quickly right now, since we are way way behind the Ma's. For example, I estimate that roughly 25% of the capabilities of Magma aren't present in Sage. Magma represents about 3 millions lines of code written over 20 years by an average of 5-8 people working fulltime (Allan Steel describes these people as their "crack team of elite programmers"). 25% of that is still a lot of work! The mission statement of Sage is to be a viable alternative to Magma, Maple, Mathematica, and Matlab, so development really just can't proceed slowly with that mission statement. Not only do we need to catch up, which is a herculean task itself, but all four of the Ma's are moving targets that are constantly improving. There are well over 2000 highly talented people employed by the Ma's, over 100 million a year in revenue. We have to catch up with and surpass that. If you know *anybody* out there who is good at programming or documentation and can help, recruit them to help on Sage. And if you write your own code and have the problem that with each new Sage release you "need to go through their own regression test suite to ensure that functionality they rely on has not been adversely impacted" then if at all possible *please* submit your regression test suite to us to include in Sage! Let us run your test suite before we make new releases. That's what the SAGE_ROOT/devel/sage/sage/tests/ directory is for. Having regression test suites run by vendors is exactly the sort of thing clients often pay for in some contexts (e.g., hardware vendors do this). With Sage, we'll do this for you totally for free, and everybody will benefit. Instead of paying $10,000/year to have us run your test suite, just do the work to submit a patch with your test suite (or get your student to do so). >>> * Users are expected to be developers >> >>What does that mean really? It doesn't seem technically meaningful to >>write "Users are expected to be developers". Expected by whom? What >>is a developer? > > I agree this is a meaningless question. Is someone who writes an Excel > macro a developer? > >>> * No glossy printed reference manual >> >>Add "You can't purchase a printed reference manual". There is a >>glossy nice pdf reference manual, but it's not for sale. > > There's nothing stopping someone from offering to print and sell it. > Yep, and evidently this is already done with the Sage docs at lulu.com (as Marshall pointed out above), though it is a bit out of date. William --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
