Hacker News new | past | comments | ask | show | jobs | submit login
Ask YC Startups: Do you use Django, Rails, PHP or Other?
26 points by wyw on Aug 8, 2009 | hide | past | favorite | 55 comments
There's been a lot of info recently about the hosting plans of YC startups but less discussion about development technology trends. I would be curious to know what the dominant technology is for developing apps among YC startups.



Rolled our own in Python.

EDIT: I should mention that CherryPy is at its heart. We love CherryPy. And our chat backend uses a secret sauce Python async framework that might or might not be open sourced very soon...


Django. You should of made a poll.


It's "should have"; the contraction is "should've".


I always thought the grammar police had weekends off.


Am I the only one who thinks that people who are correcting the grammar and spelling are actually doing a service. It really doesn't matter in a forum like this.But if you have poor grammar,you may find that the impression you make on other people may be less than stellar.A good command of the language can be incredibly valuable. [Palin vs Obama]


You're not the only one. The person who made the mistake isn't the only one who learns, either. We all do.


I don't know if it's fair or not, but I always feel like the conversation at HN is somewhat... academic, for lack of a better word. The nature of the community seems to encourage one's best behavior, or at least better behavior than I would exhibit in some other places perhaps.

Grammar correction is somewhat a conundrum. On the one hand, it IS sort of rude to point out the faults of another, especially as it can often be attributed to expedient typing. On the other hand, if I made a mistake and didn't realize, I'd REALLY REALLY want someone to point it out to me, so that I could learn from it.

Completely aside though, it's interesting to me how poor grammar seems to spread. Until about two years ago, I had honestly NEVER seen anyone use "should of" in print, and now it seems to be a prevalent mistake.


> Until about two years ago, I had honestly NEVER seen anyone use "should of" in print, and now it seems to be a prevalent mistake.

There is no way that is true:

http://groups.google.com/groups/search?as_q=&as_epq=&...


While I don't understand how that relates to the truthiness of what I've said, it looks like almost all of the results of the linked page are from 2006 forward.

Regardless, while I may be fuzzy on the timetable, it is a recent phenomena for me.


Google is ranking by date because they have the same relevance. You can play with the date ranges in the search box. I picked Jan 1 2007 as the end date because you said you "NEVER" saw this construction in text until a "about two years ago".


valid point, not sure about the example though.

Palin's speeches won't make much sense, even if they are grammatically correct and presented elegantly.


Which is why I said "command of the language". Good grammar is merely the starting point.


Linguistics study states that grammar is always evolving and there is no such "thing" as grammar. How do you know "should of" is my minor dialectical difference of English compared to yours. It's like argentinian spanish people correcting other spanish people to use vos instead of tu. Or nothing ending with a preposition. And being a good orator doesn't mean strict adherence to a one persons capture of grammar. If that was true, then shakesphere wouldn't be a good playwright!


He guessed


Django, http://schoolrack.com

We (the developers) are actually beginning to feel like it's one of the biggest Django sites on the interweb right now. Definitely not trying to toot my own horn here, but we're gaining users like mad and traffic is insane.

Disclaimer: I figure the OP's question is more directed at HN members rather than YC startups, but for the record we're not a YC startup


Sorry to call you out, but your directly quantified Quantcast results show you at less than 1k uniques per day.

There are a LOT of Django sites out there with considerably more traffic and it misrepresents the size of the Django community to suggest that, "one of the biggest Django sites" is only serving a few hundred hits a day.


We get a lot less traffic during the summer months (school isn't in session) and towards the end of the year for the holidays... but we do get quite more traffic than our quantcast page suggests.

It does look like I was totally full of my own shit though, we're clearly not that big when it comes to the other Django sites out there.


It looks like an average of 50,000 monthly uniques when school is in session. Do you mind sharing your actual numbers?

And can you hazard a guess why Quantcast is so inaccurate? It seems curious, considering you have their javascript tracking code on every page of your site. If their tracker is so broken, why use them and not Google Analytics?


Uniques monthly is totally meaningless.

This comes up with some frequency. Uniques daily * 30 != uniques monthly, after all, if you undouble on a monthly basis and you get 30K visitors each day and they're all different it looks like you have a cool 900K uniques monthly. But at the same time it means you are completely not sticky! All your users are replaced by new ones every single day.

A better measure is uniques daily, daily growth in uniques and compare that with your uniques monthly. If you're doing it right you should see (30average repeats) + (30average growth)

Those are figures that you can do some planning on. Monthly uniques does not do anything other than look good on paper. Bad websites look even better :)


As I understand it, Quantcast's monthly uniques is the totally number of different users to visit the site during that month. I assumed that daily uniques were meaningless because of the nature of SchoolRack. Not every student logs in every day. It seemed reasonable to assume that a month is a reasonable amount of time for most students, teachers, and parents to login at least once. This tells you the actual number of people who actively use such a service.


only if you put the stats script on the pages behind the login and not on the others.


Comparing Quantcast's estimates to the actual stats for a few sites I run (and have access to full logs/anayltics for) shows that Quantcast is severely (read: at least one order of magnitude in a couple cases) lowballing.


I believe the OP, I don't mind sharing my own stats, we're somewhere between 10 and 15 K uniques daily, google has us pegged at about 100k uniques (don't ask me what they're smoking there, probably something to do with subdomains), here is the alexa graph comparing ww.com and schoolrack.com, you can draw your own conclusions.

http://alexa.com/siteinfo/schoolrack.com+ww.com

A tripling of your traffic in 30 days during the 'off' season is a pretty good showing in my book, as long as it wasn't when coming from '10' users to '30' users :)


Quantcast's "Directly Quantified" stats come from the publisher embedding Quantcast's js tracking on their site - so it's kind of like making your Google Analytics data public.


just a heads up: on the about page this sentence -> Our mission from the onset has been to seamlessly connecting teachers, students, and parents through the Internet.

should be "to seamlessly connect..." ?


Thanks for the heads up, just updated the servers.


oh thank god please please PLEASE kill blackboard


http://stackoverflow.com/questions/886221/does-django-scale

Look at the first answer. It seems like there are bigger django websites out there.


Looking at the site, I have no doubts that you are doing well.

Very nice, I congratulate you. You have found a niche, and your execution is top notch.


Nice to see you gaining ground so fast during the holiday season, makes you wonder how it will go once school starts!


Pylons [ http://pylonshq.com/ ] -- a lightweight (Python) web framework emphasizing flexibility and rapid development.


Short: we use a Django frontend to an Erlang backend.

Long answer: we used Django because that's what I knew best, but we're going to switch to Pylons because it's more flexible (being a so-called "glue framework") and has less dependency on a database (tried writing a custom authentication backend in Django without a database? Good luck).

So Django is great if you have a database; less so if you don't (I guess that's the whole "full stack" thing).


Django if I use a framework, but usually straight python or perl CGI.

I'm assuming you meant "webapps".


We're not YC, but Django at Timetric.


Here we are using PHP/CakePHP on almost all our projects.

Why? I have being using PHP for almost 10 years, so it was a lot easier to switch to a PHP framework when I switched to an open source framework about 3 years ago. PHP is also easier to get on a shared hosting (maybe not anymore, but at the time I made my switch, it was a major concern).

Like : API, doc, large community.

Dislike : php syntax (->), no active record, PHP4 support (no real visibility control), deployment on windows/ISS is no easy task.

But future looks good as they plan to drop PHP4 quite soon with 1.3.

Some website we created using cake: http://www.themetropolitain.ca http://www.gogarneau.com

I hate "->".


It is funny, when every I speak php to another coder I call -> "dot". It isn't a dot, it is an arrow, but something got wired that way in my brain. "Object" "dot" "attribute"


object arrow attribute is just not as smooth


Django at Mixpanel.


Rails at Frogmetrics, PollEverywhere, Posterous


Django at DISQUS.


Rails at one; we're looking into liftweb/scala at the other.

We see the robustness of the jvm as a lot more useful at the latter- and scala seems like a good fit (better than java itself, or ruby for that matter) for the kind of datamining we're going to do. Granted, C would likely be much faster, but the ability to include third party java packages and the amount of boilerplate needed for C turned us off to that. Add in a good web framework like Lift, and we don't see the need to fracture our server-side apps into many different languages.


Django at Fluther.


Django here too...


Ooops, I should have added : not a YC startup.


Python.


ASP.NET MVC, sql server 08


Rails, heyzap.com


Django here.


Startup is Global Hostname: http://www.globalhostname.com/

DNS server is implemented in Twisted (but is not derived from twisted.names) and web site is Django+nginx. Database is Postgres.



PHP on one project and Rails on the other.


Rails + JS at Mugasha


PHP custom framework!


GWT!


django4life




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: