[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Axiom-developer] Re: Axiom Wiki FrontPage re-design

From: root
Subject: Re: [Axiom-developer] Re: Axiom Wiki FrontPage re-design
Date: Tue, 29 Nov 2005 21:24:42 -0500

> Call me old-fashioned, but in order to edit pamphlet files, I would get 
> the arch-sources and do something locally and then submit. Only for 
> small changes the web-access to the sources makes sense to me. Anyway, 
> it is not quite clear to me how modifications to the sources made over 
> the web will eventually find their way into axiom--main. Will Tim 
> incorporate them by hand? I guess the ChangeSet idea becomes a bit 
> blurred here.

ok. you're old-fashioned. but so am i so it's a bit of a struggle to
think thru the idea of hacking real sources on a website. however
bill has a good idea with this wiki site and he's managed to get
the sources online and automatically generated.

i'm all in favor of taking an idea and pushing it beyond the reasonable
and into blue-sky territory and this is one of those ideas.

the logical progression is that since we have web sources that can
be modified we should consider moving into the late 90s and stop thinking
"batch process" like CVS/Arch/Darcs and move toward "instant" processes
that would allow us to modify a web page, make axiom, and run it in a
sandbox all online.

since the website maintains a changelog/undo facility many people can
work on their local copy (a DOYEN with a browser?) and transparently
move to the wiki. 

using the wiki this way for math could be VERY cool. suppose you find
that a certain function does not compute a result. what would it take
to click on the function and be taken to it's definition (note that 
this requires tight cooperation with axiom since there are hundreds
of functions with the same name but different type signatures). and
once you've found the definition you study it for a while, write some
comments about what failed, come back in a few days with a solution,
type in the solution (and document it), remake axiom, go back to the
sandbox and test it. if it works you "promote" it to the reviewers
for inclusion into the "approved" source tree. 

each user could have their own branch of axiom.

of course, this will cause major "mainline" headaches for the "approved"
source tree but that's hardly the experimenter's problem. 

we could call this idea "scratchpad"
and make a "scratchpad" wiki where each user has their own branch.

frankly i think it's an exciting idea.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]