forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Avik Sengupta <a...@apache.org>
Subject A view from userland (was Re: What's going on here?)
Date Thu, 26 Dec 2002 16:08:22 GMT
> Forrest is *NOT* even close to be ready for prime time

Not that i disagree, but i thought you guys might want to know the
unsolicited opinion on someone who is trying to get the whole company
converted to centipede/forrest.

I've spend the last couple of weeks trying to integrate one project at
work with centipede/forrest. 

I spent about 2 full days to get the basic site generation up and
running, including some customisations in the skin. 

I have then spend a couple of days time off and on to figure out the pdf
stuff that i have sent so many mails on :)

OK, so why do i insist on using something thats so raw? Well, primarily
that forrest provides a very important functionality, and i couldnt find
any alternatives that did the job for me. For that reason i think its
worth the risk. I have been thinking about getting automated doc
generation working for about two years now. I feel its a VERY important
productivity tool,  but havent managed to implement it .. till now. 

My first efforts were at writing a few custom ant targets, but i quickly
found that its impossible to scale to multiple projects.. its almost as
much work for each new project. I've looked at anakia off and on, but
found it too difficult to use. While as the person responsible for
processes within the company, i am willing to spend time on such
projects, including debugging the tools i use, my success is totally
dependent on envangelising their use throughout the company. So scaling
(in human terms) is a key goal. And till the arrival of
centipede/forrest, i really didnt have a solution. 

(btw, i dont want to comment on maven here, except to say that i did
evaluate it)

So given the above, i am happy to have hitched a ride with forrest. What
do i need in the future? well primarily more flexible pdf generation
(dont generate for some directories, generate combined rather than
page-by-page docs for some etc).Searching, certainly. I'll probably come
up with a few more as time goes, but nothing major right now. 

And more documentation. I firmly believe that meta projects such as
centipede and forrest live and die by the strength of their
documentation. 

I dont care much about speed at the moment (tho probably i can reduce
time drastically now if i can swith off pdf generation for some
directories). Very rarely will developers use it on the command line. 
I'll more likely use it in a continuous integration settings and
automatically publish the static site. Also, i'll probably use forrest
only with centipede, and not by itself. Again, becoz i think that'll be
more scalable- people wise. 

So forrest as it stands serve's my needs pretty well (if only i can get
images in pdf :)... but i think i can crack that myself), and in fact is
my only alternative as far as i can leave. 

This mail has gotten too long already.. so i want to sign off saying..
thank you guys, great job, and keep it up. 

Regards
-
Avik






Mime
View raw message