Currently figures are attached to a fixed entry. It would be good if all figures on PM would be in a centralized figure database. This database could also contain the code creating the figure. This will make it easy to reuse figures across entries and also make small changes to figures.
It will also encourage users to take figures from PM and use them in their own texts.
I think this is a good idea for the general advantages to re-use and education on creating figures that it offers, but I'm not so sure about the benefits of re-using figures in PM articles. Which figures and which articles might be candidates for such re-use? I guess I can imagine a series of related articles that all refer to the same pictures, but in that case wouldn't you want to make one main article that contained the pictures and had all the sub-articles as attachments?
One nice thing about this proposal is that it would encourage people to include pictures in one unified way (eg. \PMfigure{} or whatever).
Maybe it could be called a "gallery" instead of a "database". --jcorneli
Direct copying of figures from one entry to another might be minimal within PM. Maybe the first figure in
could be useful in many articles. Also, a figure of the signum function could be used in the definition, and in entries related to continuity. However, if the sourcecode (not just the .eps) for the figure was available, that would increase its use a lot. With small changes the figure could have a lot of use. This would also encourage a "unified" look'n'feel on PM. It would also provide a (unique) place for people to learn from others how to create good mathematical figures. This feature would naturally create all kinds of discussions/guides about the topic.
For more involved figures the need becomes more and more important. I have spend days creating figures. Most of that time goes to learning the software.
Math is not just text and formulas. There are also a lot of standard pictures that could be useful over and over in mathematical writing. Say, transition function on a manifold, graphs of famous functions, tangent space of a surface, stereographic projection, right hand rule, etc. If I were to write lecture notes/homework solutions, it would be great to have a source where all these can be found. --matte
Absolutely, I agree with the points you're making here. (I too spent a lot of time on some figures at one point.) I still think I'd discourage reusing figures "verbatim" within the encyclopedia (e.g. instead of reusing the signum function, link to a page about it instead), but that's a minor detail: I'm definitely for the proposal. --[[jcorneli] Tue May 24 19:08:54 2005 UTC
This is something that I have commented before. Sometimes I create eps graphics using commands, perhaps using pstricks, using eukleides or even mathematica. Respectively, I add the .tex file, th .euk file or add as a comment the command used to generate the plot. Some other times I just open a graphic editor and when I'm done, save to EPS. now what abotu source code?
The thing is, EPS is a very special format. Unlike PNG, JPG, etc, it's not a binary file, it's a text file. It's a postscript file, programming code. You can open the text file see the source code. Modify it, change the graphic. It includes its own source code as documentation, which is nto the case with the other binary image formats. So I don't think that when I upload EPS files created directly (not using a translator from other language) I'm not including the source code.
The situation is similar to comparing a compiled C program, where you don't get access to the code, to a interpreted python or perl script, where the program itself is the code, you cannot get the program without the sourcecode. Think of GIF, JPG as "compiled" and EPS as "interpreted".
So, although uploading the pstricks code, the mathematica CODE, the eukleides code, the metapost code, etc is indeed a nice thing to do and Iconcurr, I don't think standalone eps files are to be shuned as lacking source code.
One could argue that "hey, but I don't know postscript, how am I supposed to change the graphic" but that argument is like claiming that distributing a perl script is not distributing its source code since somebody doesn't know perl and thus is unable to modify ithe script. – drini
I think having a figure gallery or database or whatever would be a good idea. One reason is that this would facilitate quality control. At one point there were quite a few figures on PM that were just bitmaps wrapped in EPS files, very ugly. I think the situation is better now. Another reason is that it would be easer to suggest improvements, especially if each figure is available in its original editable format. --igor
Have there been a lot of instances of need for figure reuse?
I wonder about the quality control aspect. Who will own things in the central database? How will this be managed? Seems like the current system would have to be applied to this new sort of object in addition to plain entries—which it is not ready to do at all. Thoughts? --akrowne Mon Feb 27 03:56:39 UTC 2006