Talk:Comparison of computer algebra systems

From Wikipedia, the free encyclopedia

WikiProject Systems
Systems science Portal
This article is within the scope of WikiProject Systems, which collaborates on articles related to Systems science.
Systems rating: List Class Low importance  Field: Systems
Please update this rating as the article progresses, or if the rating is inaccurate. Please also add comments to suggest improvements to the article.
Notice The old history of this article before it was merged with List of computer algebra systems can be found at /Old article history, and the old talk history can be found at /Old talk history.


Contents

[edit] Too many links

It looks like a link farm! Most other software lists don't have URLs for programs which have articles. I'll remove all these & will start to clean the external links. --Karnesky 23:57, 25 January 2006 (UTC)

oops--had accidentally rmed DoCon. Samsara added it back with the comment "slash and burn stinks." Please elaborate. -- Karnesky 01:27, 26 January 2006 (UTC)
My point is that red links are there for a reason, and that weblinks given with red links are there for a reason, too, because that's what an editor would require as a minimum to be able to make a stub; I do agree with removing external links for blue links. - Samsara contrib talk 01:43, 26 January 2006 (UTC)
Yes--I agree. DoCon's removal was a mistake & thanks for fixing it. The only other redlink I rmed was Algebra Solved!, which had no external link. I rmed URLs for bluelinked articles & the dead/foreign/off-topic URLs from External links. (I think External links still needs to be cleaned up with descriptions & possibly a few more can be removed--they are somewhat redundant.) -- Karnesky 02:09, 26 January 2006 (UTC)

[edit] Title: list vs. comparison

Is there any reason this was merged to list of computer algebra systems rather than to comparison of computer algebra systems? I'd propose moving the article, as the "comparion" is a more accurate description. --Karnesky 19:13, 26 November 2006 (UTC)

[edit] Feature comparison?

This list just compares price and license... a real comparison would include features. To put it in another way, I can use this list to quickly find out which CAS's are free, but not which of them is the best (for my purposes). Also the cyan background of the "free" cells is much too bright compared to the pastel green and red in the adjacent column. Shinobu 19:35, 14 March 2007 (UTC)

We could put another yes/no column titled "Full featured". This would be set to yes for full featured CASs like Maxima and Mathematica and set to no for specialized or incomplete CASs like Mathomatic and Fermat. - Gesslein 03:53, 17 March 2007 (UTC)

Perhaps. Although that would raise the question what the differences between various "full featured" CAS'es are. Are these programs identical? Surely not? Still, an extra column is a good idea. Maybe even more than one, if there is a desire to go into more detail. 82.139.85.207 23:59, 28 March 2007 (UTC)

I just checked this list to compare Maxima, Mathomatic and Maple, and it doesn't help much. Useful comparisons might include: maximum digits precision, (or "highest integer"); some kind of benchmark, say for a big factorial or maybe factoring a given big number; number of functions; families of functions. (trig, calc, primes, etc.); postfix or infix notation; maximum variables per expression; highest number base; graphing (y/n); extensible; callable from other languages or programs; etc. --AC 06:19, 6 September 2007 (UTC)

[edit] Color coding is POV

IMO having the green background color for open source and the red background color for non open source cells, subconsciously suggests that open source is better. Green typically suggests good when compared to red suggesting bad. Regardless of whether or not it is, wikipedia shouldn't come out on one side or the other. —Preceding unsigned comment added by 66.193.5.99 (talk) 17:03, 8 November 2007 (UTC)

I agree (in fact came to the talk page to say the same thing) guiltyspark 21:49, 15 November 2007 (UTC)
The same complaint is current at Talk:Comparison_of_statistical_packages. JonMcLoone (talk) 16:36, 28 April 2008 (UTC)
We had this discussion at Template talk:Yes. The consensus was that green means yes & red means no & that we aren't prescribing a value judgment. The 'but yes' and 'but no' templates were deleted for this very reason. --Karnesky (talk) 17:44, 28 April 2008 (UTC)

[edit] Please add MATLAB Symbolic Toolbox

It is one of the most popular CAS systems. —Preceding unsigned comment added by 71.37.58.190 (talk) 05:39, 26 January 2008 (UTC)

[edit] Encalc a CAS?

I don't think Encalc should be listed here, it is not a system that can be run locally and it doesn't appear to do any algebra. It is only a units conversion web service as far as I can tell. --George Gesslein II (talk) 10:52, 31 May 2008 (UTC)

I agree. Online is OK, but a units converter is not a CAS. I will remove T68492 (talk) 11:16, 5 June 2008 (UTC)
Good job, thanks! -- George Gesslein II (talk) 00:28, 6 June 2008 (UTC)

[edit] Supported platforms

I think that the supported platforms table should indicate whether compiled executables are available or whether you have to compile it yourself. For people who do not know how to compile code, systems that do not provide pre-compiled versions, remain inaccessible.

Either a (*) leading to a note, or perhaps such systems should be marked as "No" with a note "Source code should compile for this system " —Preceding unsigned comment added by T68492 (talk • contribs) 11:14, 5 June 2008 (UTC)

Do you have an example of a tool and platform that you'd like this to be the case for? I think there is too much grey area for that to be really valuable. Which distributions of Linux, BSD, and Unix would you choose to represent the answers for the rest of the column? What if something is only compiled for a particular version of windows or OS X? For some cases, using the "depends" template with a note might be appropriate (e.g. relies on cygwin on windows or wine on *nix). I wouldn't necessarily count out non-binaries as "unsupported;" sometimes binaries aren't distributed, but there remains active user assistance (commercial support or otherwise). --Karnesky (talk) 01:20, 6 June 2008 (UTC)
Probably quite a few of them. But lets look at Maxima as an example. If you look at the download pages... [1] and [2] you see that there is a Windows installer (.exe) and the same for Linux (I assume thats what a .rpm file is) but for other platforms you get only the source code that you must compile. So as a Windows user I can download and run Maxima with no special software or skill, but as a Mac user I must get a compiler and know how to use it. I think that makes Windows a "Yes" and Mac a "Yes*" (or "No*" depending on whether we think the majority of WP readers can compile software or not). I think the question of what flavor of an OS is supported is a different question. I think we have to interpret the column in terms of the most important flavors. eg If it fails to work on XP either as a pre-compiled .exe or when you compile it yourself then Windows should be a "No" but if works on XP but not on 95 then it might be a "Yes". If it works on XP when you compile it yourself but the .exe only works on 95 then that would be a "Yes*" (or "No*") Vista should probably not yet be required, but soon should be. Likewise for the main Linux platforms and Mac OSX Intel vs OS9 or OSX classic.
Personally I would take the tougher interpretation. Most people know nothing about compiling software, so those without installers should be "No*" where "*" is labelled as "The source code supplied will compile for this platform if you have an appropriate compiler" T68492 (talk) 09:38, 6 June 2008 (UTC)
Probably quite a few of them. But lets look at Maxima as an example. If you look at the download pages... [3] and [4] you see that there is a Windows installer (.exe) and the same for Linux (I assume thats what a .rpm file is) but for other platforms you get only the source code that you must compile. So as a Windows user I can download and run Maxima with no special software or skill, but as a Mac user I must get a compiler and know how to use it. I think that makes Windows a "Yes" and Mac a "Yes*" (or "No*" depending on whether we think the majority of WP readers can compile software or not). I think the question of what flavor of an OS is supported is a different question. I think we have to interpret the column in terms of the most important flavors. eg If it fails to work on XP either as a pre-compiled .exe or when you compile it yourself then Windows should be a "No" but if works on XP but not on 95 then it might be a "Yes". If it works on XP when you compile it yourself but the .exe only works on 95 then that would be a "Yes*" (or "No*") Vista should probably not yet be required, but soon should be. Likewise for the main Linux platforms and Mac OSX Intel vs OS9 or OSX classic.
Personally I would take the tougher interpretation. Most people know nothing about compiling software, so those without installers should be "No*" where "*" is labelled as "The source code supplied will compile for this platform if you have an appropriate compiler" T68492 (talk) 09:38, 6 June 2008 (UTC)
Not all Linux distributions support RPM packages out-of-the-box, though. So, as a Linux user, you MIGHT be able to use the binary RPMs, but you might not. Further: on OS X, you can get maxima binaries using fink. It is also part of SAGE, which there are binary OS X downloads of. The mailing list for maxima has plenty of posts from OS X users--I fail to see how it isn't a "supported" platform. Other products will have similar potential exceptions & you haven't proposed to do with any program that is distributed only as source. I think the current tables are consistent with other software comparison pages. Before any change, we should make sure that we address some of the "gray area" that will be common for anything marked "yes*" or "no*" under your proposal. --Karnesky (talk) 13:02, 6 June 2008 (UTC)
If you can get a MacOSX .dmg file then that is a "yes", even if it is odd that you have to get it from someone other than the project. Perhaps a better example is YACAS. Looking at the download page [5] there is ONLY source code. It gets ticks in every platform when you can't actually get it ready to use on any platform.
You are right that there are lots of gray areas (different flavors of an OS, different hardware, extra software that some systems, amount of configuration, compiling, code hacking you need to do etc) and in the end, the page for that system is the place to put all the detailed requirements and caveats. But I think the basic test should be "Can a TYPICAL (insert OS) user go and get/buy (insert system) that will usually WORK without alteration" and I think on that test having to have a compiler and knowledge of using compilers before they can get it to "work", is at least significant enough to highlight.T68492 (talk) 15:47, 6 June 2008 (UTC)
yacas is a great example. It'd border on the ridiculous to say it was "unsupported" on all platforms. There are third-party binaries available & yacas makes efforts to ensure their C++ version is cross-platform & explicitly claim support on the platforms listed in that table. To say otherwise would seem to be projecting an bias against source distributions.
There is no NPOV definition of a "typical" linux/bsd/unix that I am aware of. I wouldn't consider compilation "alteration."
The table is accurately labeled "operating system support" and has an informative note that the only criteria used is whether it runs without emulation on a platform. I see no advantage of changing the heading and description to say that binaries are available & I don't know whether or not this would warrant yet another table. --Karnesky (talk) 16:55, 6 June 2008 (UTC)
I have nothing against source distribution. That seems good. But ONLY source distribution is worthy of mentioning. I expect that the large majority of people that might be able to make use of YACAS are not equipped to be able to use in in source-only form. There is no download from the project of a Windows version of the current release of YACAS. I think this page is misleading those people but not telling them. At the very least the definition above the table should warn them that that this table does not imply binaries are available and that they may need to compile it themselves from whatever language the system is written is. T68492 (talk) 10:37, 9 June 2008 (UTC)