First work: study
OsvaldoTCF
- Scenario
- Scope
- Available technologies: tools for design and documentation,
frameworks, components, database (to be used during development), scm,
etc.
- Vote
Thank you all, Merry Christmas and happy new year.
stootch
I. Let's keep it simple from the underlying database's structure's
point of view.
1. flat structure like: master table holding availabe sub-forums,
detail-level-one table with threads, and detail-level-two table with
posts. Perhaps both detail tables can be combined, but that depends on
what we want to store/display/achieve. Posts don't need to be further
nested into tree-like structure thus simple 'ORDER BY
DATE_TIME_POSTED' should be sufficient.
2. from time to time I spot a forum where administration can create
sub-forums within sub-forums within... and so on. IMHO this is kind of
nightmarish both for users and RDBM, but we should decide it from the
beginning, as it requires completely different database layout. I vote
'no' :-)
II. Advanced templating can be introduced at a later stage. With CSS
solely one can make real magic without changes in HTML.
III. Most of our efforts should be spent on (1) security and (2)
robustness.
1. SQL injection, cross-scripting, session-capture... Further,
deeper analysis is definitely required.
2. Whether the forum works as CGI of FCGI it should make as few
calls to database as possible. I know this sounds obvious, but I saw
some projects killing DB engines down, and they were designed by
professionals.
Can we make use of AJAX? Sure, but not necessarily to beautify our app
or make it dancing. As for now one idea comes into my mind: let users'
browsers do some parsing eg. replacing smileys with icons, or even
processing bbcode-like entries (they'll surely be required). Needs
consideration and testing.
IV. OpenIDs and/or regular registrations? Since Google incorporated
OpenID I've observed growing popularity of this authentication
mechanism.
V. Almost every time I look at source HTML of the sites I visit, I see
ugly formatted content. I mean *really* ugly. No matter if the given
page survives validation, it's almost unreadable for human. With no or
little effort our app can produce beautiful HTML/CSS/JS and that WILL
be noticed by professionals. And professionals are the guys others
listen to ;-)
For now I see nothing unique about this project, maybe except being
written in Pascal. So let's show something interesting to the
wilderness: mirroring. Any change commited into *any* running forum
(e.g. new user registration, new post etc.) would be propagated to all
existing mirrors. Perhaps this feature would allow mirroring of whole
forum or chosen sub-forums. One mirror dies for any reason, remaining
continue to work as if nothing happened.
Mirroring will most certainly require special API as simple
replication on database level won't do the trick. Besides, hosting
companies don't allow their customers do replication unless they get
generously paid.
Any other ideas that would make this project really unique? Remember,
as we want to gain public interest in Pascal we need to show something
really interesting.
Stootch
Graeme Geldenhuys
>
> IV. OpenIDs and/or regular registrations? Since Google incorporated
> OpenID I've observed growing popularity of this authentication
> mechanism.
So add to this, integration with Gravatar (http://www.gravatar.com/)
would be a nice touch as well. It uses the email address of the poster
and pulls in a avatar image (if one is available). No need to store
images in the user preferences for the forum.
Regards,
- Graeme -