[fpc-other] Re: Offer to repair and maintain the FPC community website (repeat msg, no HTML)

[Tomas Hajny]

Hi Pete,

Alright, here we go. Sorry, you'll realize soon that I can write 
longer e-mails than you although or maybe because English is not my 
native language... ;-) Note that all comments below (listed in 
completely random order) express my purely personal view (including 
my own impressions from previous discussions on various topics in the 
FPC core team) rather than any kind of official position of the FPC 
team. Note also that I have no privileged access to any of the 
servers used by FPC.

For start - I understand your points regarding number of clicks, easy 
to reach download pages, and impressions our site may bring to 
newcomers and I agree at least some of your comments and 
recommendations. From this point of view I personally appreciate your 
interest in improving the situation and offer to help with that 
directly.

Having stated that, I believe that it's useful to clarify some 
reasons for the current situation - not as excuses for this or that 
existing issue (and especially not to try excusing dead links or 
improperly working search function), but rather to provide background 
information on certain priorities which were considered when building 
this site (some of them may be equally valid as before, some of them 
may be seen as lower priority nowadays):

- Up to now, the whole site runs on a private machine without server-
grade guaranteed connectivity, etc. Providing simple means of 
mirroring the main pages (including files offered for download), 
etc., has always been a high priority. This mostly means static pages 
for the main part without any PHP, database back-ends, etc., because 
these all require special configuration on the mirroring servers. 
Obviously, this applies to the main pages (and content for download) 
only, not to the bug tracker, wiki, forum, etc.

- As of now, the content of the static pages is managed on our SVN 
server in the form of simplified HTML pages (the full pages are 
regenerated in case changes in SVN on daily bases, or the refresh may 
be enforced manually). This translation copes with things like 
changes of download URLs for mirrors, etc. While not very user-
friendly for traditional content editors, this approach is fairly 
easy for programmers dealing with SVN regularly anyway (read: the FPC 
core team) who are not interested in spending a lot of time in 
configuring special web content management systems, learning how to 
work with them and how to protect them from attackers, etc.

- As a nice side effect of the current approach focusing static pages 
and simple content, it's worth mentioning that the pages load 
extremely quickly regardless of connection speed and used device 
(small size, static, no complex constructs).

- It's worth reiterating that FPC is a volunteer based activity. 
Dependency on a single person able to perform certain administration 
tasks is an issue. This is actually visible on the community fora 
(supported by a special software which is understood by just one 
person from the core team who hasn't been even so active with FPC 
recently) which occassionally suffer from downtimes, some smaller 
issues which have never been fixed, etc. The simpler the tools are 
and the lower effort necessary for managing them the better.

- The site has been created mostly as support resource for existing 
users. You're completely right that attracting new users is important 
and we should support this goal better. Nevertheless, most accesses 
would probably always come from people who are not completely new to 
FPC. This means that links to 'more advanced' resources like viewvc 
shouldn't be seen as something inappropriate (in spite of the fact 
that such stuff may be 'scary' for newcomers). Actually, viewvc 
provides much better visibility on "what's going on" and how much the 
project is really alive than any kind of forum, because it shows the 
real progress and changes of the codebase rather than just questions 
and answers. Admittedly, this certainly wouldn't be obvious to 
newcomers without providing them with some background first.

- Mailing lists, NNTP groups and Web fora all have their advantages 
and disadvantages. Choices are influenced by combination of personal 
taste, working habits, frequency of posts, resilience (remember that 
we have no enterprise-grade servers with high availability), options 
and alternatives for SPAM protection, etc.

- The FPC core team values its independence. There are no 
advertisements, neither from the FPC team nor from any external 
provider of hosting services, etc.

- It's worth mentioning that even the dependency on SourceForge which 
provides us with means for much wider and more efficient distribution 
of load generated by release downloads is a bit problematic because 
it increases efforts necessary for the release processes (due to more 
complex procedures for publishing of files) and it also imposes 
certain restrictions to us.

- Providing single click download for approximately 20 different 
distribution packages (even considering just the latest official 
version) may not be an easy task, especially not for static pages. 
Doing so while also considering increased server loads after releases 
(which requires distribution of the download traffic to mirrors) is 
even more difficult. Single distribution package of everything (all 
hardware architecture, operating systems, etc.) makes no sense.

- The community fora support internationalization (both for the site 
and its controls and also for the content). Some of our users value 
this option apparently.

- Dead links are obviously the easiest thing to solve. You can easily 
start immediately by reporting them through our bug tracker.

- If the FPC bug tracker doesn't work properly with Chrome - well, 
bad luck, but Mantis bug tracker is a fairly common and reasonably 
advanced bug tracking solution, so if there are issues with a certain 
browser (or its particular version), it may be better to report such 
issues directly to the Mantis developers using 
http://www.mantisbt.org and/or to developers of that browser.

- Using a Wiki page may be useful to prepare a more complete concept 
regarding how things could or should work in the future. Directly 
creating an independent alternative version of the site (as mentioned 
in one of the previous posts) may be an option too, but obviously 
with a higher risk of burnt efforts.

Note that all of the above is meant to help you not to miss possibly 
important aspects while thinking about the new concept. There's no 
need to argue this or that bit (you don't need to convince me about 
anything) - but I'm obviously willing to (try to) clarify anything 
possibly unclear.

Tomas