[uf-dev] Status/Updates...

[Ryan King]

On Nov 29, 2006, at 5:01 AM, Brian Suda wrote:

> I did manage to get my HG access working just fine and i have
> committed some updated test files. I am now working on getting all the
> XSLTs into some logical place to then also commit them. (that should
> be soon).

Awesome.

> My XSLT files need some debugging. I have managed to get them to work
> just fine with XSLTPROC locally, but then when i use PHP's
> XSLTTransform() things don't always work. Then i also tried them with
> the W3C's new XSLT2 service and they don't work at all[1]. So once i
> get things committed, if anyone has more experience with XSLT2 i might
> need some help chasing-up bugs - i get an error[2]:

I've been thinking for awhile about setting up some automated testing.

I was inspired by a comment by DanC (maybe blog post), where he  
mentioned that the OWL working group setup automated testing and  
published the results on the web (eat-the-dogfood style). Let's  
brainstorm some ways for us to do this.

If we can setup a distributed system, it would be very easy for  
people to automate testing on a bunch of xslt implementations.

> ...
>
> Also, there is an interesting article about Creating Passionate
> Users[3]. It tries to talk about why jargon is a good thing...
> basically, the points that got me were that they argued NOT to
> dumb-down to newbies, but to stick with the jargon and aim higher.
> There has been/is a long thread about splitting/creating mailing lists
> for newbies, or restating things better for new comers to the lists. I
> know this is the DEV list, but i'd rather spend my time having
> conversations here, than on the discuss list at the moment. Thoughts?

2 things:

1. The splitting is not just a matter of having a newbie list vs. an  
advanced list. It's also about splitting the two types of  
conversations (using microformats and researching/discovering new  
microformats) up a bit so that people only interested in the former  
aren't overwhelmed by the volume.

2. There's nothing more influential than running code (in this context).

-ryan

--
Ryan King
ryan at technorati.com