[dev] IMP login template...

[Marko Djukic]

Quoting Eric Rostetter <eric.rostetter at physics.utexas.edu>:

> I'm with Jan (et al).  We need to look at each page separately.  Does
> it improve (or worsen) the look?  Does it make the coding easier? Does
> performance improve, decrease, or even matter for that page?  I think it
> is a page by page thing.  Or at least a module by module thing if you want
> consistent code within a module.  But I think it really should be page by
> page...

a few thoughts fwiw...

1. just from a brief look at nuno's patch, although all credit to nuno i think
he's doing a good job tackling a not so easy task, i believe the patch needs
quite a bit more cleaning/optimisation before we can say for sure what's the
loss/gain by switching to templates. there seem to be too many tags being set,
which could be the biggest performance loss.

2. i don't believe that templating is something to be applied only to underage
apps, so to call them. in fact, i don't see another solution that can give apps
rapid customisation and deployment. from a pure business viewpoint it's a *lot*
cheaper and quicker to have one graphic designer crank out a set of html
templates than to have a php coder hack through code/html to get the same
result. of course if we do continue to get performance losses of 15% with
perfectly written/optimised code, then it's time to relook at the templating.

3. the last point is how relative is the gain/loss. for example how noticeable
is a 5% performance loss, if horde could jump yet a few more steps ahead of the
competition, boast a template driven webmail, that a client could have their
own look in no time, a webmail that could be viewed perfectly in a browser, on
a PDA, on a phone? i believe imp is the best webmail out there and is what got
me started on horde and believing in horde. you guys have done a great job
getting it to what it is both in terms of features and performance (sure beats
that stuff on yahoo i have to use once in a while!). but there's also a bigger
picture to look at, apart from optimising the existing code to death.
in fact it was talking to two different clients of mine this week that got
imp/horde the biggest compliments i heard so far - apart from these two
companies suddenly having an amazing webmail after 4-5 different ones over
several years, they have discovered it is accessible both on PDAs and cell
phones (their primary use of internet since their employees are 75% on the
road). accessible yes, but layout-wise terrible (100px width x 200px height
cellphone screen), but with a template background that could be fixed in a
snap.
so the point is, looking forward, what would get users/clients jumping up and
down more: saying "we have lobbed off another 1% off the performance time" or
"you can now have perfect webmail to all your employees wherever they are and
whatever they use to log on". with all the boom in GPRS, 3rd generation phones,
PDAs that i see in italy, and probably in rest of europe, i think it's the
latter which will get more jumps.

4. could a solution to not messing up the stability/performance of present day
imp be to branch it off and work on a template-only driven version of imp?

marko