Re: New Boost Website: Open to Volunteers!

From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2022-04-16 21:30:30

On 16/04/2022 21:48, Vinnie Falco via Boost wrote:
> On Sat, Apr 16, 2022 at 12:39 PM Stefan Seefeld via Boost
> <boost_at_[hidden]> wrote:
>> making the new site more modular in its way to syndicate content
>> coming from individual boost projects
> 
> We are adding a database

I would be opposed to any form of website which relies on a database, 
unless it is SQLite.

I say this from bitter experience. Databases go wrong without 
maintenance and we've seen on many occasions in the past people happy to 
volunteer to do stuff for Boost for many years until they quietly 
disappear, then the thing limps on but becoming ever increasingly 
erratic and problematic, and then we have to go get people to start 
ringing other people and then we find said service was being run off 
some academic server and nobody ever knew about it and then academic 
people start accusing Boost of network intrusion etc etc.

We could just avoid repeating all those past mistakes by not designing 
in a need for a database. I remind everybody here of my past travails 
with Boost's pre-github wiki and issue tracker. Let's not repeat that.

Where my ideal would be at is a static site generator such as Hugo. 
Which is like what we've already got, except several generations newer 
and better technology. It can syndicate content from feeds just fine, 
you stick it on a five minute cronjob, it's designed to work well doing 
that.

I'm not against databases, so long as it's a SQLite file and said file 
can get deleted at any time and it doesn't really matter. I used to use 
one for website visitor counters, as an example.

I also think a github repo which acts as the content for the website 
would be a wise thing to retain. People can raise PRs, they can get 
approved, content appears on the website. I think that a good pattern to 
retain rather than yet another portal for reviewing and checking content.

Niall