[Analysis] Templates: The good, the bad, the ugly

[Myk Melez]

J. Paul Reed wrote:

>On 13 Dec 2004 at 15:17:08, Myk Melez arranged the bits on my disk to say:
>
>  
>
>>Sure, but could we have used something better?  I don't think so.
>>    
>>
>
>Yet, you really don't provide any justification for that answer.
>

To be honest, I don't remember the details of my research, but I 
remember that TT beat the alternatives hands down at the time on the 
factors I considered, and I'll wager it would do so again if we did the 
research today, probably on feature set, customizability, code quality, 
community/support, and project vitality.

>I was having a conversation with Christopher Hicks about this, and I
>provided a lighter weight solution which I think is better.
>
>Of course, I didn't bother mentioning it here because a) I wrote it, so "of
>course I think it's better" and b) I wrote it long after TT was chosen.
>
>Personally, I think TT is a huge pile of crap, if for no other reason than
>it's slow.
>
"It's slow" is vague; how is it too slow for you?  It's fast enough for 
me today, including on the very busy b.m.o (which slows down sometimes 
for other reasons but loads templatized pages quickly otherwise), and 
it'll be much faster in the future under mod_cgi and SpeedyCGI.

For what other reasons is it "crap"?

>That doesn't mean that we shouldn't do a post-mortem and be honest about
>where the policy decisions, implementation efforts and the technology
>failed. We should be similarly honest about the successes.
>  
>

Agreed.

>Also, the claim "We shouldn't look at any other solutions now because the
>one we picked years ago was the best one at the time" isn't compelling. I
>don't know if you're making that claim, but I've heard it made, and that's
>just wrong.
>  
>
I agree, and I'm not making that claim, because if a better solution is 
available, and its long-terms benefits outweigh the cost of switching, 
we should do so.  I do think that we're unlikely to find a better 
overall alternative, not even considering the migration cost, but I'm 
willing to be convinced otherwise.

>Me, I didn't find using TT "pleasurable" at all. I found it's loop control
>particularly stupid.
>  
>

What's stupid about it?

>>This is actually the right thing to do in some cases, particularly when
>>processing takes a while and we can give the user updates on its
>>progress, f.e. when sending mail or modifying multiple bugs.  Of course,
>>we would need iterative template output for that, but that can be hacked
>>into TT2 and comes native in TT3, an alpha of which was recently
>>released.
>>    
>>
>
>Ahh... the software salesman's mantra: "It's in the next version."
>  
>

Actually, it's available now.  I hacked it into TT2 for a different page 
on b.m.o years ago, and it could be used just as easily for pages for 
which iterative status updates are useful.

>Either way, that way it's done now is ugly, and problematic because of
>include ordering.
>  
>

What's the include ordering problem?

-myk

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bugzilla.org/pipermail/developers/attachments/20041213/bc570935/attachment.html>