| Time |
S |
Nick |
Message |
| 00:14 |
|
pdurbin |
interesting anti-Django rant by prologic: http://irclogs.shortcircuit.net.au/%23circuits/2015-02-23.log.html#t22:06:27 :) |
| 00:15 |
|
prologic |
lol |
| 00:15 |
|
prologic |
not quite anti-Django rant |
| 00:15 |
|
prologic |
but it's not suitable for all use cases |
| 00:15 |
|
prologic |
:) |
| 00:15 |
|
pdurbin |
nothing is :) |
| 00:16 |
|
prologic |
to be fair; I'm a bit frustrated by the fact I've been porting a ~150k codebase form Django 1.2 to 1.4 (yes still old and soon to be supported) and the volume of hacks, overrides and monkey patches just makes using Django very annoying and really what we've done is created our own framework and just so happen to use some bits and pieces of Django (1.2) |
| 00:17 |
|
pdurbin |
too tough to go straight to 1.8 I assume |
| 00:20 |
|
prologic |
no that's true - nothing is really |
| 00:20 |
|
prologic |
well my colleagues obviously feel so |
| 00:20 |
|
prologic |
so I'm limited to 1.4 at this time |
| 00:20 |
|
pdurbin |
gotcha |
| 00:21 |
|
prologic |
but with this kind of developmment attitude we'll always be behind the eight ball |
| 00:21 |
|
prologic |
"so to speak" |
| 00:21 |
|
pdurbin |
well, sounds like a big undertaking. that many lines of code |
| 00:21 |
|
prologic |
but that's business for you |
| 00:21 |
|
prologic |
yeah it is |
| 00:21 |
|
prologic |
previous attempts have taken an elapsed 3 years of time |
| 00:21 |
|
prologic |
(not necessarily effort) |
| 00:21 |
|
prologic |
by pretty much all the backend/devops devs |
| 00:21 |
|
prologic |
and now I've been on it for 3+ months now |
| 00:21 |
|
prologic |
hopefully I'll finish what was started |
| 00:22 |
|
pdurbin |
that's the spirit |
| 00:22 |
|
prologic |
yeah :) |
| 00:22 |
|
prologic |
look at the bright side eh? :) |
| 00:22 |
|
prologic |
but tbqh some of Django's own code and even ours just pissed me off |
| 00:22 |
|
prologic |
:) |
| 00:23 |
|
pdurbin |
searchbot: lucky wtf per minute code |
| 00:23 |
|
searchbot |
pdurbin: http://blog.codinghorror.com/whos-your-coding-buddy/ |
| 00:23 |
|
prologic |
and our documentation is well err lacking :) |
| 00:23 |
|
pdurbin |
http://www.osnews.com/images/comics/wtfm.jpg |
| 00:24 |
|
prologic |
yeah |
| 00:24 |
|
prologic |
peer review/ code review |
| 00:24 |
|
prologic |
is kinda very important to any good software |
| 00:24 |
|
prologic |
really |
| 00:24 |
|
prologic |
haha |
| 00:24 |
|
prologic |
I'll share that with my colleagues |
| 00:24 |
|
prologic |
might get a laugh or two :) |
| 00:24 |
|
pdurbin |
:) |
| 00:25 |
|
prologic |
nice one |
| 00:27 |
|
pdurbin |
prologic: did I ever have you glance at the Google Doc linked from https://github.com/IQSS/plaid ? |
| 00:27 |
|
pdurbin |
the first goal: A web framework that support rapid application development |
| 00:28 |
|
pdurbin |
the first item listed under technology stack: Django |
| 00:30 |
|
prologic |
ouch |
| 00:30 |
|
pdurbin |
well |
| 00:30 |
|
pdurbin |
so far it's going fine |
| 00:31 |
|
prologic |
depending on what you're protoityping |
| 00:31 |
|
prologic |
I wouldn't use Django |
| 00:31 |
|
pdurbin |
here's a nice little app that was built quickly: http://roadmap.datascience.iq.harvard.edu/ |
| 00:31 |
|
prologic |
so I guess your use-cases are "fine" :) |
| 00:31 |
|
prologic |
*nods* |
| 00:31 |
|
prologic |
that looks Djangoy :) |
| 00:33 |
|
pdurbin |
I'm not saying it's fancy. |
| 00:34 |
|
pdurbin |
but it went from idea to in production quickly |
| 00:34 |
|
pdurbin |
not that I had anything to do with it :) |
| 00:36 |
|
prologic |
*nods* |
| 00:36 |
|
prologic |
which is totally fine |
| 00:36 |
|
prologic |
bunch of templates (looks bootstrapy) |
| 00:36 |
|
prologic |
a few views |
| 00:36 |
|
prologic |
some models |
| 00:36 |
|
prologic |
and hey voila |
| 00:37 |
|
prologic |
I guess the thing is if you were to create the same thing in say circuits.web |
| 00:37 |
|
prologic |
the process would almost be the same |
| 00:37 |
|
prologic |
just |
| 00:37 |
|
prologic |
pick a templating engine |
| 00:37 |
|
prologic |
pick an orm |
| 00:37 |
|
prologic |
write those up |
| 00:37 |
|
prologic |
then the rest is the same |
| 00:37 |
|
prologic |
and voila |
| 00:37 |
|
prologic |
s/write/wire |
| 00:39 |
|
pdurbin |
I'm not even sure what templating engine and ORM are used by Django. |
| 00:40 |
|
pdurbin |
I think the templates might be a Django-specific thing. Dunno if you can use them with other frameworks. |
| 00:41 |
|
prologic |
they use their own |
| 00:41 |
|
prologic |
they basically wrote their own |
| 00:41 |
|
prologic |
Django's templating engine and syntax is actually inspired by and very similar to Jinaja2 |
| 00:41 |
|
prologic |
and their ORM is very similar to SA too |
| 00:42 |
|
prologic |
and no |
| 00:42 |
|
prologic |
if you use Django -- you're stuck with Django |
| 00:42 |
|
prologic |
*really* |
| 00:42 |
|
prologic |
it's a very hard coupling |
| 00:42 |
|
prologic |
once you start to do things like (like we've done) write many hundreds of templates, many hundreds of models, (worse) write business logic in your models and views |
| 00:43 |
|
prologic |
it's quite difficulty to uncouple that then |
| 00:43 |
|
prologic |
you could "port" the templates easily enough with some effort |
| 00:43 |
|
prologic |
but views, models and business logic would not be trivial |
| 00:49 |
|
pdurbin |
there's hard coupling to the ORM too? |
| 01:26 |
|
prologic |
yes I think so |
| 01:26 |
|
prologic |
it would not be trivial to swap out Django's ORM for another I don't htink |
| 01:26 |
|
prologic |
I'm not sure exactly how compatible Django's ORM is to any other |
| 01:26 |
|
prologic |
e.g: peewee, SQLAlchemcy, SQLObject, etc |
| 01:26 |
|
prologic |
they're all a bit different |
| 01:34 |
|
pdurbin |
gotcha |
| 01:38 |
|
pdurbin |
so if you wanted to maximize your ability to switch frameworks (a somewhat dubious goal) you should choose popular components like jinja2 and sqlalchemy, both of which are supported by django, I think |
| 01:39 |
|
prologic |
yes they are |
| 01:39 |
|
prologic |
by various extensions |
| 01:40 |
|
prologic |
I guess it's really abot locking yourself in |
| 01:40 |
|
pdurbin |
yeah |
| 01:40 |
|
prologic |
but yeah I agree - one should not choose a framework with a goal in mind of then swithcing away form it later on |
| 01:40 |
|
prologic |
but it's important to understand the tight coupling and lockins obviously |
| 01:41 |
|
prologic |
and that if Django doesn't do exactly what you want your choses are to either find a way to extend it, get it patched upstream |
| 01:41 |
|
prologic |
or monkeypatch it yourself |
| 01:41 |
|
pdurbin |
you should say, "I'm gonna write 150 thousand lines of code in this thing. I'd better like it!" |
| 01:41 |
|
prologic |
I always detest the later option |
| 01:41 |
|
prologic |
yeah |
| 01:41 |
|
prologic |
well if we had any idea of the complexity of the business requirements up front |
| 01:41 |
|
prologic |
I guess we wouldn't have picked DJango |
| 01:41 |
|
prologic |
dunno |
| 01:41 |
|
prologic |
I guess that's evolutionary software development for you though |
| 01:41 |
|
pdurbin |
you could do worse, I suspect |
| 01:42 |
|
prologic |
something to be said about waterfall sometimes |
| 01:42 |
|
prologic |
and the careful design process |
| 01:42 |
|
prologic |
probably :) |
| 10:05 |
|
|
aditsu joined #sourcefu |
| 13:41 |
|
|
aditsu joined #sourcefu |
| 16:04 |
|
|
rruma joined #sourcefu |
| 16:10 |
|
|
aditsu joined #sourcefu |
| 19:20 |
|
pdurbin |
nice: [ANNOUNCE] Apache Solr 5.0.0 and Reference Guide for Solr 5.0 released - http://mail-archives.us.apache.org/mod_mbox/www-announce/201502.mbox/%3CCAKiERN5Zx83qJSLOUp1qTYOnfN4h1_zVo=2LFuE=7Li5fknZVg@mail.gmail.com%3E |
| 19:22 |
|
|
aditsu joined #sourcefu |
