Hi Ted,
On Feb 22, 4:16 am, "Ted Kosan" <[EMAIL PROTECTED]> wrote: > On Thu, Feb 21, 2008 at 8:30 PM, mabshoff <[EMAIL PROTECTED]> wrote: > > sage-devel - the idea is that sage-edu rolls up a number of nice > > (initially optional) spkgs that get merged into Sage as they mature. > > The same applies to any extra code written by sage-edu. sage-edu > > should function as a test bed for various bits and pieces of code like > > the interactive widgets, which the research people have little > > interest in. I doubt "you guys" want to do your own releases, so sage- > > edu works as upstream, somewhat like currently the sage-combinat > > people do. > > It is my opinion that sage-edu should be a separate entity from > sage-devel and that the main connection between the code that sage-edu > creates/distributes and the Sage computation engine should be through > a well-designed API similar to what Mathematica has: > > http://www.outbacksoftware.com/mathematica/mathlink-jlink.html sage-edu is not meant as a separate entity, but as specialized group to discuss education specific aspects of Sage development. The volume of sage-devel can be so high that many people prefer not to subscribe to it. Having something will better signal to noise ratio for that specific groups of people can only be good. > Since Sage takes inspiration from Mathematica in so many other areas, > my thought is that using mathlink as a starting point for what a Sage > API should look like would make sense too. I think that API is just ludicrously heavy and way to complicated. Sage development has always been guided by KISS and the above diagram is far from that ideal. > Now to spkgs. As the person who created the JMOL-src package, I don't > want to have the 3rd party applets that sage-edu makes available to be > forced to be distributed in source code form. Beyond this, sage-edu > might eventually make dozens of education-oriented applets available > and, if sage-devel does not want to deal with an education-oriented > applet like GeoGebra, I don't see why it would have any interest in > including education-oriented applets in Sage at all. Well, in the end there ought to be only one distribution - there is no point in forking the codebase. It is not that the "research people" don't want to do the work, but that the "edu people" need to get organized. Doing that on sage-devel hasn't worked too well in the past, not William has suggested to do it that way. > If sage-edu is a separate entity from sage-devel, it will also need to > have its own source code repository and releases. Well, that is certainly possible in form of their own corner in the spkg repo and its own upstream hg repo, but the goal is to be upstream and aggregate patches and not to start diverging in code base. > > > 2) Would sage-devel be willing to expose a standard API that can be > > > used to access the Sage calculation engine? > > > I doubt that is really needed. If somebody wants to implement some > > mathml abstraction layer (as Joel suggested below) that would be fine, > > What is needed is something like Mathematica's mathlink, not a mathml > layer. Sage needs an API like this for the same reason that > Mathematica needs one. We certainly disagree in that one. I also don't see the need for Mathematica to have such an API, but in the end we will agree to disagree on that one. > > but while in the past William has thought about splitting off the > > notebook as its own project I am not seeing that happening any time > > soon. > > Why not? The notebook's developers have stated multiple times that > they would much rather do math-oriented development instead of work on > the notebook. In the mean time, the current number of notebook > tickets has reached 70. Sure, most of them enhancement requests. What the notebook developers expressed is a desire to do more math oriented development and that has caused the notebook to not advance as quickly as it might have had if more people were pushing it. But it isn't like there is a bunch of people sitting on the sideline waiting for the "research people" to give up on the notebook so they can swoop in and show us how it is done properly. Having done most of the releases for the last four month I just have to tell you that I seriously doubt that the "edu people" can duplicate the effort that has gone into making regular releases. If there are people interested in improving the notebook it needs to happen in the Sage code base, no "if, else or buts". Maintaining more than one code base is a nightmare and I can assure you that due to the funding the "research people" are getting (MSR as well as "other") the vast amount of code written will come from that group. Any kind of duplication is wasteful and in the end the needs of both groups should be met in the same code base. What also needs to happen is that somebody starts pushing patches important to the "edu crowd" in trac to be reviewed and merged and also tickets about the notebook to be elevated in priority and fixed. So far there is no such person from the "edu crowd", one could even go so far that there is next to none from that camp that actually writes [serious amounts of] code. Somebody like Jason Grout has his feet in both camps, I mean somebody who is purely interested in Sage as a teaching tool. In the very end I need to quote Linus: Talk is cheap, show me the code. As long as this is all planning and hand waving I see little benefit from having this discussion. As you mentioned yourself there is plenty of work in trac to do. So far nobody else has jumped in on your side and volunteer his or her free time and energy to make things happen. It would be nice to know if anybody wants to sign on and get things done, which is what we all want to happen to make Sage a better tool for the research as well as edu crowd. Sage should be flexible enough to accommodate both camps and in the end we should have much more in common than separates us. > Ted Cheers, Michael --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
