emacs info obsolete


On Feb 20, 9:00 am, Jens Thiele wrote:
> Xah Lee writes:
> > no time. The simple problem is that nobody uses it, excerpt maybe just
> > 100 core GNU fanatics around the world.
> I don't consider myself a fanatic and still I really like reference
> documentation in info format. The info browser in emacs is really nice -
> searching and navigating the documentation is fast and easy.
> jens

i too, find emacs info system fantastic. As far as i know, better than any other integrated system i'm familiar with. (possibly except Mathematica. Though, i know that for example Macintosh CodeWarior, ThinkPascal, etc IDE has simular features that are comparable, but i haven't used these much to confidently remark)

However, in the context of this thread, i don't support info, or support advancing it in anyway. Yes, it is great in emacs, but basically it is 1980's technology. Far advanced at its time, but today, it is fairly inferior comparative to other IDE's system such as Mathematica, or whatever is today's version of MCW, or perhaps Microsoft's VisualStudio)

For example, in other systems, it can include graphics, syntax coloring, multiple panes or tabs in the help system. Though, these are probably technically possible with info, but basically is not done and incompatible with info's nature.

as another example, many today's IDE integrated help system are based on HTML/XML and associated tech (javascript, DOM, css, etc). With these, its technical power is quite beyond what info can do. Such power include, for example, integration with the web, usable in browser too, post-processing with hundreds of tools, dynamic generation with hundreds of tools, interactivity with js or visualbasic and other langs etc. (just look at google's various services, such as gmail, blogger, google code, igoogle, google map etc.)

as another example, info's source is based on texinfo, with effectively one single tool. For html based ones, there are literally few hundred times more tools, in wide variety of languages, or user level applications, and few THOUSANDS times more developers.

texinfo and info, is basically outdated in many ways, technical and social, that it cannot support itself into the future. Though, it works great today still, but in coming years, if emacs insists on sticking with info today, it'll pose a problem. (as another example, unix man pages, which info and FSF intend to replace, are today basically already obsolete)

I would suggest, that emacs developers start to embrace html/xml, which is standardized system in the industry. Put aside the cultish thought that emacs info is superior, which ceased to be true since perhaps 2000.

emacs already has lots of packages that deals with html/xml. Muse, planner, org, all deal with documentation in some mark-down format that are oriented with html. Then there's nxml-mode which is today one of the most advanced system for xml, written by the well-known xml expert James Clark. There's also js2-mode, and ejacs by Steve Yegge, which is also very advanced. Each of these is over 10k lines in elisp.

if emacs dev put aside the “emacs way” mentality, there are many opportunities to advance emacs forward in major ways.

I'm not saying that we should ditch info or right away, just not insist on it. Embracing the above tech can easily pave a way that naturally replace info's role in emacs.

∑ http://xahlee.org/


No comments:

Post a Comment