[isf-wifidog] WiFiDog Roadmap, first draft

[Francois Proulx]

The problem I see, is that since none of us ( at Ile Sans Fil ) will be using the MySQL schema on a day-to-day basis, it'll be a pain to keep it in sync with the developpment. I'm repeating myself, but let's make it clear for everybody, I am willing to help anyone who's interesting in using it daily and wants to maintain the schema. The initial schema is the real bottleneck, once is done, most queries are not using highly complex SQL, it should not be a problem to run them on MySQL.

Will somebody craving for MySQL support step up on this mailing list and assert that he will maintain the schema ? I will give the momentum by giving the part of the schema I converted and helping him (or her ? who knows).

Until this happens, that'll be status quo for me. It's seems obvious that if we start to support MySQL we need to keep it working....

--- Mina Naguib <webmaster at topfx.com> wrote:

From: Mina Naguib <webmaster at topfx.com>
Date: Wed, 08 Jun 2005 10:06:05 -0400
To: WiFiDog Captive Portal <wifidog at listes.ilesansfil.org>
Subject: Re: [isf-wifidog] WiFiDog Roadmap, first draft


On 6-Jun-05, at 3:32 PM, Matthew Asham wrote:


> Benoit, thank you for the post, this makes your vision very clear to
> understand and I fundamentally agree with you.
>
> So far as MySQL vs. Postgres goes, I greatly prefer MySQL for general
> database work but Postgres provides the functionality we need to build
> on top of.  I personally believe the MySQL support should be removed
> entirely.
>
> Matthew
>

I cannot say I agree with this opinion.

When people say "free database" MySQL comes to mind, not PostgreSQL.

Yes, I know that pgsql is better, yes I know it has more features,  
and yes, I've been eyeing it for more serious database work ever  
since I started using it in this project.

However, It is NOT the de-facto db that is available to most people.

Our early rationale for convincing ourselves that it was a good  
choice was that people who are interested in running a WifiDog auth  
server are highly experienced in the technical nature and would  
probably run their own server with whatever software they need;  
However, that assumption has been proven wrong quite clearly by  
people who want to use our software along with the mainstream open  
source db.  At least one person who posted to the mailing list passed  
off on our otherwise excellent software simply because of our  
slightly eccentric DB choice.

We need to make a decision and make it clear.  The "DB abstraction"  
that makes it "hard to add a second DB layer" is not a good DB  
abstraction.

If we decide that, hey, we will only do PGSQL then remove all partial  
mysql support and post that decision in bold letters in the FAQ.   
While we're at it let's maximize the use of pgsql.

If not, let's move all "cleverness" away from the DB and into the PHP  
code so we may get away with 1-syntax-fits-all for maximum db support.