HNHacker News
TopNewBestAskShowJobs

psranga

561 karma · joined February 20, 2007

submissionscomments
psranga··on Twitter Inventor About To Launch His Next Project, Code-named Squirrel
IIRC, Twitter was started as a way to send personal bits of info to people you knew. Why else do they have the tagline "What are you doing?".

That idea didn't work. Hence I would definitely agree that a lot of luck resulted in Twitter being repurposed as an broadcaster of impersonal information to people who you don't know. (which is pretty much the opposite of the original intent).

psranga··on California sales taxes collection drop 50.9%, income taxes 43.6%
Can you elaborate why you say getting into debt "is the right thing to do"? Why do you think it's unreasonable to reduce spending when tax collection goes down?

Personally, I find the parent's commonsensical prescription appealing. The parent probably meant was that if govt gets into debt, "higher and more ubiquitous taxes" will have to collected in the future to pay for the debt.

psranga··on Ask HN: Explain AWS
Yes, does anybody have data on latency and throughput of EBS "disks"?
psranga··on Can you trust 37signals with your password?
Protect yourself against lax companies like this with the PwdHash Firefox extension. http://www.pwdhash.com/
psranga··on Can you trust 37signals with your password?
Just like Microsoft. :)
psranga··on 6 Line EventMachine Bugfix = 2x faster GC, 13x requests/sec
My comments were about "big" self-contained projects for which it's worth it to have their own internal infrastructure and idioms etc. An example would be a compiler or an interpreter or a web server.

For smallish programs like what you describe, I agree that manual memory management can be decent overhead.

Many "big" projects try to reduce the memory management hassles by embedding an interpreted language such as Tcl or Lua for implementing non-performance intensive tasks.

I'm not advocating writing everything in C++, but with the right tools and in the right circumstances, the benefits of C++ can outweigh the annoyance of manual memory management.

Another reason I'm skeptical of garbage collection is because I read that unless you do things in a specific way you can have references hanging around in Java resulting in memory bloat.

psranga··on 6 Line EventMachine Bugfix = 2x faster GC, 13x requests/sec
If we interpret what cperciva said to mean 100% correct proofs you're correct. But you can do a lot of things in C++ to avoid memory leaks/reading from freed memory etc.

Some of them are:

1) are auto-freeing templatized heap object class (the compiler will automatically call destructor when objects goes out of scope, call free in the destructor).

2) Overload free/delete to fill up freed memory with 0xdeadbeef etc so that code that reads from there is likely to crash due to invalid data.

3) Use one of the non-libc memory managers which will track memory allocs and frees and tell at exit if all memory was freed. Some will even have tricks to tell you which allocation was not freed.

4) If all else fails, use Valgrind for the really tough ones.

Coupled with a good regression/unit test suite, memory leak tracking will be an insignificant part of your development effort.

psranga··on Ask HN: Review my new single service website
Very nice. How about if the result appeared inline below the "Lookup" button instead of on a new page?
psranga··on Ask HN: Selling the software without the business
A common mistake people good at something make is to assume that the average person who does that is also as good as them. In this case, they are competing against what seems like an established enterprise software company. It's unlikely such a company has the environment to (a) attract good developers and/or (b) ensure good developers produce at peak output.

But if the technical aspects of the tool matter to customers, then it's possible that suits will see the value in replacing their technology with something better. Because from a business point of view, it's FAR less risky to buy than to develop.

I'm not saying these guys will be able to sell their technology alone (customers are very important), but it's highly non-obvious to me why their technology is worthless.

psranga··on Ask HN: Selling the software without the business
I know of at least one instance of the opposite. This team sold a technology to an established software business for a decent money.

If it took these guys three years getting to where they are, it will take at least that much for any competitor/established player. Three years has to be worth something.

If you do get acquired for the technology, expect to be have a payments spread out over multiple years (~5 yrs) with employment at acquirer required to collect.

I agree with edw that having even a single beta customer will dramatically increase the $ that an acquirer will pay (Humans trust others' judgement more than their own).

psranga··on Confessions of an Entrepreneur's Wife
Several points in no particular order why I think this may be a "good" article (i.e., it's good storytelling), but I'm still critical of this article.

While writing this reply, I skimmed a couple of food startup companies on Wikipedia. It looks like these guys tried to do a SoBe (started 1996 bought by Pepsi in 2000; they year they started their company) and failed. I'd be very interested in an account of the SoBE story.

1. No reasoning in the article about why they thought this drink would sell.

2. First expenses are to buy two Land Rovers? AFAIK, these are not the cheapest SUVs. And this is literally like in a recent bad movie I watched ("Role Models").

3. On the same lines, they got an RV and two salespeople to drive around nationwide in an RV handing out samples? Was this drink even available nationwide? I've never seen it. Why were they spreading themselves so thin?

4. No mention in the article about how they got things done on the cheap. Which is a staple of all startup stories. Looks like they splurged on everything.

5. I had to read the BevNet link somebody posted to this article to find out that this drink costs much more than sodas because it uses real juice. It costs $1 to make and sells for $1.29. Is there a tiny probability that this could have contributed to the failure of the drink?

6. I mean this thing costs in the range of Red Bull and Starbucks but they market it as a competitior to Pepsi and Coke?

7. As I kept reading the article, I noticed how everything was being blamed on everybody else other than Bill. Bill did no wrong, and others let him down. I did not see enough evidence to convince me that this was the case.

8. Though billed as wife's story, this article is equally about Bill. So I have a problem that Bill is written up from the point of view of an adoring wife.

9. The BevNet thread shows that there are some people who are very angry with Bill. Btw "Wayne" (he doesn't get the courtesy of his real name being using, I think he's Mike Gilbert from the bevnet thread) has a book out to put his side of the story. This article sounds like an attempt to rehabilitate Bill's reputation.

10. Notice how they heavily criticize LKCM for "attempting" to bankrupt the company. But I'm supposed to find nothing suspicious about Bill proposing a deal with a new investor which would give the new guys 51% of the company (meaning management control with ability to replace Bill etc).

11. In short, at the end of the article, I was completely unsurprised at the total failure of the company. There is only one lesson to learned from this story for me: if hire a Bill, you'll fail. But I already knew that.

psranga··on Confessions of an Entrepreneur's Wife
I was bothered by the lack of introspection in the article.
psranga··on Git vs Mercurial: Google Code Analysis
I bet Google picked Mercurial because it was easier for them to hack on and adapt to their infrastructure. A "libgit" is on lots of people's wish list, but the official developers seem pretty happy with the hodgepodge of binaries.
psranga··on Writing Software is Like ... Writing
A closer analogy imho is with lawyers.

Writing doesn't necessarily involve logical thinking of the type programming involves, and it doesn't seem to involve "hacking" (in the best sense of the term), and it doesn't involve working withing constraints in a pre-existing system.

psranga··on Google is replacing conference room paper schedules with solar-powered LCD's, updated via 802.15
Maybe I'm dense, but if you go to the wrong room you'll only know you're at the wrong room. How do you know what the correct room is?
psranga··on Coding Horror: Is Open Source Experience Overrated?
Usually companies try to evaluate a candidate independently; unless your side project is exactly what the company wants, I'd be surprised if your open source projects are the deciding factor in your getting the job.

But your o-s contribs can get you an _interview_ and indirectly help you in the interview.

Hopefully all practice you've been getting through your side projects would have made you a better developer and that will be apparent in the interviews.

From this point of view, I would say major work building one's own app makes one a better developer than minor work on a major project.

Your skills development is what matters in the end, not the "name recognition" of having contributed to such-and-such project.

psranga··on Coding Horror: Is Open Source Experience Overrated?
You can do everything you said without contributing to an open source project.
psranga··on Coding Horror: Is Open Source Experience Overrated?
Yes. The best developers I know do not contribute to open source projects, and have no interest in doing so.

I can understand why. It take a HUGE amount of effort to digest a big codebase; why do it when the most fun in programming is to be had when you're designing/implementing a new system?

psranga··on Facebook rejects funding at $4 billion valuation, may not raise more
I'm a little puzzled by the skepticism about Facebook's utility. I don't know about you guys, but losing touch with old friends is something I feel bad about all the time. Facebook let's you keep informed in a shallow way about people who you don't meet often (enough). That's certainly better than being uninformed.

I would actually pay a small amount for something like Facebook.

If you build something with utility, usually it's possible to derive value.

psranga··on Ask YC: Negotiating a job offer
Walking away doesn't work if all companies are the same (and they are).

As soon as you start negotiating, they'll know you got the info from your friend (these HR people are not dummies even though they don't code :)). Don't tell them though; given them plausible deniability. Say something idiotic like you looked at cost of living and think that this is too low or something; just give them a excuse to up their offer.

Since they like hiring in pairs, you have some pretty good leverage. Go for it, you'll probably succeed.

I learned the hard way that all salary offers are open to negotiation (even if the person making the offer claims that it's not).

psranga··on The Education Visa: It’s Everywhere We Need It To Be
The book Illusions of Entrepreneurship claims that existing companies create more jobs than new companies. Starts at page 153: Here's the google books link: http://bit.ly/m4vXV
psranga··on The Founder Visa
Others have said it, but I'll say it again.

Maybe in 1989, this would have made sense. But capitalism has had enough converts that most countries have pretty low barriers to high tech startups now.

IMHO, people who are already entreprenurial in their home countries will be the last people who will want to immigrate to the US. Maybe pg looks at the number of companies founded by immigrants and thinks that he can increase that number by creating a new visa class for founders. Immigrants start businesses because American culture is much more stable, open to startups, tolerant of failure, and gives people second chances; so many immigrants let go of their previous cultural biases and become founders.

If you're already an entrepreneur in a poorer country like India and China, why would you immigrate from there to a country where all expenses are vastly higher. I mean, American companies are setting up development centers in India and China. For richer countries like France and Germany, you give you any existing networks and go to a new place. Seems like a net loss to me.

You have stay in a culture for a while before you recognize business opportunities there. I bet that most of the immigrants who start things do so after staying in the US for 3+ years. My claim is that a new immigrant will do worse at business than average.

You'll have to let in lots of smart people and hope that some start companies. Maybe you can set up the rules so that it's easier for the smart new arrivals to start a business as soon as they're ready.

Currently, due to various rules, the most common route to permanent residency takes approx 6 years of continuous employment with the same employer for an immigrant to gain unrestricted rights to do business (this requires permanent residency). So for somebody who arrived for a Master's degree (being simplistic and using Master's as a proxy for smartness), it will be 8 years after arrival. This same person may have understood the culture well enough in 3 years and be ready to start a business.

psranga··on The Founder Visa
You rank applicants to the visa program by Erdos numbers? Is this the same Erdos number as http://en.wikipedia.org/wiki/Erd%C5%91s_number? Or do you rank investors?

In either case, you're ranking people based on WHO they know, not what they know. Worst kind of ranking possible (it keeps competent outsiders out).

psranga··on The Founder Visa
A quick search shows no mention of Illusions of Entrepreurship. Excellent book, btw.

That book claims that existing companies create more jobs than new companies. Starts at page 153: Here's the google books link: http://bit.ly/m4vXV

psranga··on Cybersecurity Act of 2009 to allow warrantless searches of all personal email
I don't see anything about searching personal email here. Is the original arcticle inaccurate?
psranga··on The 48 Laws of Power
I bought this book a few years ago (5 IIRC). I started skimming it and got pretty sad about the things one supposedly had to do to make it. Then I lost the book a few days later when I left it on a flight.

I took it as a sign; I didn't bother going to go to lost and found or buy another copy. I wonder if I did the right thing. :)

psranga··on How My Brain Kept Me from Co-Founding YouTube
Hmm, maybe I'm too retro for this guy, but 16 bytes for an element in a linked list does sound excessive. :)
psranga··on Turn Off Your Step-Thru Debugger
When you're first working on a large chunk of unfamiliar code, stepping in gdb is an AMAZING way to learn how the code works.

If this guy is advocating adding prints to code, then his edit-compile-debug cycle must be very fast (i.e., smallish programs). On my large-scale project, it's much faster for me to restart the program with new breakpoints than compile and run again.

The latest gdb has checkpoints. That is a significant advantage over printf-based methods.

Of course, reasoning about programs is a very very effective tactic to solve bugs. If you write you code so that you can reason about it, you reduce the need for both printf's and debugs. As somebody pointed out above, this usually means writing short methods.

psranga··on Happy 15th Birthday to Tcl and the Tk Toolkit
Tcl is an amazing language. Very underrated. It's "EIAS" (everything is a string) philosophy, and the duality of strings and lists is very very powerful. It's got one of the best-documented and precisely written standard libraries I've ever seen (and the "extended standard library" tcllib is of similar quality).

E.g., if some script returns "350 seconds to execute command <more stuff here>" you can hold your nose and just treat the string as a list and extract the 0th element :) ('lindex $return 0').

It's speed leaves much to be desired. My ideal stack would in a large project would be:

1. Tcl shell on top 2. ... talking to either Python or C/C++ below

psranga··on Blueprint: A CSS Framework
When I last looked at Blueprint, it looked pretty good for brochureware websites. But when I tried to use it for something else, I quickly ran into lots of limitations. I ended up coding my own CSS.

What do others think?

We really need a programmer-friendly abstraction layer over the mess that is CSS.

← PreviousPage 7 of 8Next →