keiretsu

i agree. i'm using a module system myself and it's way better than OO

The issue with Rails for me wasn't the database bottleneck. It was the ActiveRecord. It was using far too much memory and was taking up the bulk of the page loading time. After i migrated the app to my php framework, I downgraded to from a 2GBz 1GB RAM server to a 2GHz 512MB RAM server and i still get a performance boost. Running the rails app on the 512MB server was painfully slow.

Writing your own framework is a good compromise if it can give you that performance boost that the other frameworks can't give you. I've tried the other PHP frameworks and the best page loading time i got was 0.9s.

Don't assume that everyone who uses PHP don't use a framework. Anyone who uses a PHP framework can get similar development speed as someone who uses Django/Rails.

There's a reason why smart people uses PHP instead of Rails: Most notably it's proven to scale (Yahoo, Facebook). I have personally tried Rails and love it to bits. But the performance issue is not worth it. In the end, i wrote my own stripped down Rails-like framework in PHP in 1 week and migrated the app in a few days. Performance boost was very noticeable. (Avg page loading time dropped from 1.5 secs to 0.3s)

And to those anti-PHP fascists, yes it's possible to write neat PHP code. There is no scattered SQL in my app since i follow Rails' MVC structure. And since I code "functionally" and don't use PHP OO features, so its half-baked object system doesn't bother me.

I don't mean pure functional. I mean i use lots of functions instead of object methods. I don't like mixing data with functions.

so i would assume that you do not hold hastily put together teams with less regards as long as the thing they're building is not lame?

i agree. It is usually people problem. Personality clashes etc.

i hate OO programming because of inheritance. It pissed me off to have to look through layers and layers of parents to search for a method. Stick with good ol' functions for easy debugging.

I'm seeing a trend of hackers looking for fellow hackers online to apply for the Ycombinator program. I was wondering how effective a partnership would such an arrangement yield?

2 dudes. Total oblivion to each other's quirks. Joining force to undertake a significant project. How realistic is that?

Also, would Ycombinator have less regards for such hastily put-together teams, irrespective of their skills/ideas?

erm. referring to web startups here.

And yes. You heard me right. I've talked to a number of so called internet biz dev "pros" and they actually advocate outsourcing site development to offshore places and concentrate on marketing. I was like, WTF? Now i know why the successful web startups all originate from hackers. These internet biz dev "pros" just don't get it

The most uncomfortable thing is that my non-techie co-founder actually agrees with this thought. Maybe it's time for me to find another cofounder.

You can really tell the diff between an outsourced site and a non-outsourced site. An outsourced site will most likely be .aspx (India's fav), have a rigid design/structure. Eg. http://scriptovia.com

An in-house version usually has its own unique design/identity Eg. scribd.com

i think most of the business fogeys don't understand the importance of having your own in house technical team. these fogeys actually think outsourcing development to offshore teams is the best way to go.

Right. Technology is the lifeblood of most successful web startups. So when was it good to outsource your lifeblood to a bunch of unknowns miles away?

amen

But what if you have already done the site, got 1000+ users, got some press etc.

Yeah. Ok. you haven't got incorporated yet, but based on the man hours spent, isn't it unfair to give the late-coming guy 50%?

then start one BEFORE you get married

Sure. It's damn cool that you can create your wonderful product at a cost of only $1000. But should you reveal the truth to the VC? Or should you lie to them that it cost $50,000? The reason is if you tell the truth, the VC might think since your cost is so low, the valuation of the company they'll give would be low too

I'm using the same logic as a buyer knowing the cost of a supplier. If you were told the goods cost $1, the likelihood of you bidding $100 is lower than if you were told the goods cost $80.

speech-to-text. i don't see how high a barrier of entry is this. bmw and the other manufacturers would most likely incorporate this in their electronic systems if it becomes popular.

Now, i am sure some of you have encountered a good-for-nothing co-founder who seems like a good fit early on but turns out to be a disappointment as time dictates. Assuming you've decided enough is enough and you decided to look for a new co-founder, what should be your course of action. Note that your existing co-founder has some equity in your startup.

Should you start a new company with the new co-founder replicating the old startup (which should take about a month at most) or stick with the old company?

P.S. I'm assuming a web startup here. So replication here means doing a similar site.

Who is talking about respect here? I'm talking about biz models!

Most web 2.0 co. are based on advertising revenue.

this is an effing retarded counterpoint. Boy are their biz model scalable to hundreds of millions.

I have yet to see any true ipo-able web business to appear so far. When i mean truly ipo-able, i mean it has a REAL business model and one that can scale to billions. Something like Paypal or Ebay. Let's face it, Youtube, Facebook are all toys. They have eyeballs but no scalable business model. The best hope i've seen so far is Prosper.com. It doesn't have the massive eyeballs yet but hey, at least their biz model is scalable.

Does this mean that all the good billion-dollar ipo-able web ideas have been taken up? Is this it?

so what's new?

Assuming you just can't get any traction (more than 10,000 users) after a year, would you give up?

a few more iterations and you'll end up with optimusprime

stick with this name. Auctopus. Alright! Where's my 5% share of the company now?

er. that's the point of this topic. i'm asking for tests to narrow down the pool of interviewees and screen out the bad ones so i do not have to waste my time to interview them. why wouldn't they want to join my company when i'm going to pay them 50% above market rate?

it's called probability. It just depends on how you define a good employee and a bad employee and based on that definition, you can prepare a hiring program and test if that program works.

Almost all successful companies have some sort of verifiable hiring programs. facebook has their puzzles as a screener test. It is not a universal test but it helps. If you are able to solve the puzzles, there should be a higher probability that you are a better employee to the company than someone who can't solve the puzzles.

All they need to do is verify if those puzzles are good at screening good hires. If yes, well done then. They've got themselves a great hiring tool.

I'm not asking for the one almighty test that can screen if an employee is good or bad. I'm asking for tests that can help in increasing the probability of hiring good employees. For eg. for a sales job, you absolutely have to do a personality test. If the test indicate that you are introverted and have bad people skills, that you are not likely to be a good hire. Is the personality test a universal test? No, it's not. But it helps.

ok fine. if iq tests are such a bad idea, what verifiable tests are there that can be used to classify a good employee and a bad employee? I'm sorry but gut feel just doesn't cut it.

hiring based on "feel" is like trading based on "feel". My back hurts. Damn, short this stock. Oh god. I feel alive whenever i see this stock. Buy!

it's always good to base your hiring on something that is testable. Call me crazy but hiring is a long-term continuous process. It's best to have a program that's testable and verifiable, rather than simply be based on gut feel.

the gist is to check if the site is accurate

see my comment below. I'm not planning to ask them to take a full blown iq tests. Just those "iq games" that i find are empirically good in classifying the smart from the non-smart

so you are inferring that there is enough statistical evidence to prove the hypothesis that a high iq person does not make for a good employee.

yes. i will talk to them, give them problems and see how they respond. but iq test is still useful. by the way, i'm not asking them to do those full blown iq test. Just those fun-and-games "test" like http://www.lumosity.com. This site really does work. I've test it out on my friends. The smart ones get high score. The dumb... err... less bright ones get average scores.