xirium

Its a theorectical average salary. You've got a very large chance of earning nothing or almost nothing and a very small change of being a zillionaire. You get one or more attempts at getting a good outcome. The average outcome would be approximately 100,000 pounds.

> those skill sets lend themselves too many opportunities

If you're in the UK and you've got fair qualifications in maths, computing and physics then your choices are as follows:

- Finance - 60,000 pounds per year. Insane hours. Very little job satisfaction.

- Consulting - 50,000 pounds per year. Long hours. Moderate job satisfication.

- Film post processing - 40,000 pounds per year. Long hours. See your own work at the cinema.

- Application programmer - 30,000 pounds per year.

- Teacher - One year training course, possibility of 40,000 pounds in five years if you can handle unruly kids. (Many don't make it.)

- Games programmer - Insane hours. Project may require porting code or using a legacy engine. Awful third party libraries licensed to shorten development cycle. Pay depends on previous success. May require beginning as a game tester at minimum wage.

- Startup Programmer - Theorectical average salary of 100,000 pounds per year. Large control of project specification. Good chance of recognition and job satisfaction.

> game development now lends to larger teams, less individualism...

In the summer of 1996, I got some work experience testing a game at Bullfrog. In 1996, Bullfrog games were taking between 20-22 man years to develop. As a lead programmer, the latter phase of the project would consist of 12-15 hour days of bug fixes and build failures. In the subsequent 12 years, I haven't had the inclination to apply for a job in the games industry.

As a stopgap measure, check the Paul Graham "facts" ( http://news.ycombinator.com/item?id=60357 ) and the Bruce Schneier "facts" ( http://geekz.co.uk/schneierfacts/ ).

The constraint to be optimised is code size, not readability or speed. A strongly-typed language would be at a disadvantage if casting between datatypes was required.

From the article: Based on the original perl golf, Code Golf allows you to show off your code-fu by trying to solve coding problems using the least number of keystrokes.

I'll be a pedant and note that Code Golf is based on Perl Golf, which is based on the obfuscated Perl contests, which are based on the IOCCC. The major differences are implementation language, specified problem and a hard limit on submission size.

Two weeks ago, the BBC had an article about the shortage of mathematicians and physicists who persue teaching ( http://news.bbc.co.uk/1/hi/education/7433339.stm ). Would you want the best mathematicians and physicists writing games or teaching the next generation of students? Currently, there's a shortfall in both cases and the reason is quite simple: the pay and conditions are atrocious.

> min-width which is not supported in all browsers

There's a very simple workaround. Reference a transparent, one pixel image and specify the desired minimum width.

> The community experience is the driving force behind reddit/digg/HN/mixx/slashdot/etc successes, not the technology.

Slashdot only went open source after they gained a lead with their own instance of the code. Slashdot only went open source after increasing demand from its community. Since going open source, Slashdot has subsequently gained bug fixes and a much more scalable implementation at reduced cost. However, I cannot name any other instances or derivatives of the code which are active - not even Squishdot Zope clones.

So, if you've got a big archive of posts and a large number of active users then open sourcing your code only entrenches your position by reducing costs.

From the article: outside very poor African countries, Aids is confined to "high-risk groups"

To me, this indicates that without an expensive preventive programme, a worldwide epidemic would have occurred.

I hate to make the comparison due to the difference in significance, however, there were similar sentiments about Year2000 fixing. The few remaining problems were dismissed as isolated incidences and people wondered why so much money was spent.

Layering stateful and stateless protocols minimises abstraction leakiness ( http://www.joelonsoftware.com/articles/LeakyAbstractions.htm... ) while maximising future re-use.

Unfortunately, I cannot recommend any books on the subject. People who write successful protocols spend their time maintaining successful implementations. People who write books about protocols have typically implemented protocols which are used within small niches, if at all.

If you want to learn by example then the best suggestion I can give is to read RFC821 (email delivery) ( http://www.faqs.org/rfcs/rfc821.html ), RFC822 (email headers) ( http://www.faqs.org/rfcs/rfc822.html ), RFC1094 (NFSv1) ( http://www.faqs.org/rfcs/rfc1094.html ), RFC1813 (NFSv3) ( http://www.faqs.org/rfcs/rfc1813.html ), RFC3010 (NFSv4) ( http://www.faqs.org/rfcs/rfc3010.html ), RFC1436 (Gopher protocol) ( http://www.faqs.org/rfcs/rfc1436.html ), RFC1459 (IRC protocol) ( http://www.faqs.org/rfcs/rfc1459.html ), RFC1945 (HTTP/1.0) ( http://www.faqs.org/rfcs/rfc1945.html ), RFC2616 (HTTP/1.1) ( http://www.faqs.org/rfcs/rfc2616.html ) and supporting documents for a mix of text and binary protocol specifications, reasonings and evolutions.

If you're not accustomed to reading RFCs then they may seem quite terse. Persevere and it will get easier after the first five or so. At the very least, read RFC822 because it is the basis of web and email headers.

I only said that you was probably a bot. A good method to determine if you are a bot is by drawing attention to the situation. A bot would very probably ignore the implication. So, you're either a very well written bot or a person. I now think that you're probably a person.

In a previous post ( http://news.ycombinator.com/item?id=214742 ), you describe circumstances in which you evaluate contributions based upon the contributor.

Possibly. In the article, he mentions that he has one million lines of Java in production.

Receiving email via Postfix and qmail has previously been described on this forum ( http://news.ycombinator.com/item?id=186879 ). To send email, pipe data to /usr/lib/sendmail or suchlike. Postfix, qmail and other MTAs typically install a wrapper at /usr/lib/sendmail

However, if you're hosting on ADSL, the IP address of your mail server is probably blacklisted, so you may have to configure your MTA to forward messages to your ISP's MTA. I found this easiest to configure in Sendmail, which can be configured to co-exist with other MTAs.

> Sometimes I think all I do is write code to convert data from one type to another.

You'd be good at ETL.

> Bonus!!! The Unknown or exprired link when posting to Hacker News after typing for a long time.

Definitely. I hit it after reading your comment. If you ever want a technique to discourage long and insightful comments, this is it.

Implementing stateless protocols over stateful protocols over stateless protocols is powerful and minimises the leakiness of your abstraction. Interface (stateful) over HTTP (stateless) over TCP/IP (stateful) over packets (stateless) is desirable.

> that's how I feel when programming web apps.

HTTP is extremely verbose with 0.4KB headers in each direction and JavaScript was originally a domain specific language to implement features within Netscape3.0. Neither are optimised for current usage.

We could make spam uneconomic with a similar implementation.

And did you know that you're probably replying to a bot that posts dozens of stories from alleyinsider.com ( http://news.ycombinator.com/item?id=204760 http://news.ycombinator.com/user?id=fromedome http://news.ycombinator.com/submitted?id=fromedome http://www.ycombinator.search.xirium.com/cgi-bin/search.cgi?... ), nowhere else, and has never contributed to a discussion on this forum?

Well, how does it work if people postdate copyright?

Is Clean Room Software Development orthogonal to Test Driven Development or not? Clean Room minimises tests; Test Driven maximises tests. Clean Room minimises side-effects when code is written; Test Driven minimises side-effects when code is changed.

That's a sneaky trick with plenty of scope for abuse. If I postdate the copyright on a book by five years then, after 20 years or so, it will be harder to determine the true copyright expiration date.

I'm not a lawyer but I've got a countermeasure for this type of legal abuse. Copyright is intended to be for a fixed term but I don't recall any details that allow or prevent this period to be deferred. Therefore, if a book has next year's copyright then it may not yet be in copyright.

RIAs? Rich Internet Applications? Stick with JavaScript. Flash and Flex aren't currently supported on the iPhone or 64 bit browsers. Microsoft Silverlight is Windows only, which will arguably make your software inaccessable to a large pool of early adopters.

Recent developments with SquirrelFish will make Safari's JavaScript implementation 10 times more efficient and subsequent developments could provide an additional factor of four improvement. Firefox is likely to follow with a similar implementation.

The History Of Oil by Robert Newman ( http://video.google.com/videoplay?docid=-5267640865741878159 ) also mentions the possibility of a US only depression.

Its easier to learn programming when you're a kid. Its easier to learn natural languages when you're a kid. How long would it take an adult to learn a second natural language to conversational fluency? Some people can do it in six months or less. Some will never achieve it.

no one will sue you for patent infringement till you have money, and once you have money, people will sue you whether they have grounds to or not. So I advise fatalism. Don't waste your time worrying about patent infringement. -- http://www.paulgraham.com/softwarepatents.html

We do advise the companies we fund to apply for patents -- http://www.paulgraham.com/softwarepatents.html

You're doing it wrong if it becomes a grudge match. See http://blog.guykawasaki.com/2006/03/the_art_of_driv.html

(Yes, I am aware that Guy Kawasaki marketed the Apple MacIntosh at launch.)

> every business dislikes competition.

Not true. A struggling business dislikes competition. A healthy business likes competition. A competitor validates your market and frames a choice. With a competitor, a customer doesn't just have the choice of buying from A or not. A customer now has to consider buying from A or B (or buying nothing). When you've got dozens of competitors, you've got a market.

A submitted article has high visibility and has no risk to karma. Contributing to a discussion has low visibility and risks karma. An RSS reader and a bookmarklet makes article submission semi-automated but insightful discussion takes more effort. Perhaps there should be more points for discussion.

The next most easy option is to delay submissions by an hour or so. If two or more users submit the same URL in one hour then no-one gets credit. This would ensure that two or more bots monitoring the same blog would cancel each other. More generally, it would encourage more diverse content to be submitted.