User: Password:
|
|
Subscribe / Log in / New account

Quotes of the week

Improved code and documentation quality – we’ve observed that the peer pressure from 'Social Coding' has driven engineers to make sure code is clean and well structured, documentation is useful and up to date. What we’ve learned is that a component may be 'Good enough for running in production, but not good enough for Github'.
Netflix

The Internet was done so well that most people think of it as a natural resource like the Pacific Ocean, rather than something that was man-made. When was the last time a technology with a scale like that was so error-free? The Web, in comparison, is a joke. The Web was done by amateurs.
Alan Kay
(Log in to post comments)

Quotes of the week

Posted Jul 19, 2012 6:13 UTC (Thu) by butlerm (guest, #13312) [Link]

> The Web, in comparison, is a joke. The Web was done by amateurs.

The way I see it, there are only two major problems with the web. One is that a modern web browser is about a thousand times more complicated than the requisite TCP stack, so you have all sorts of version skewing and feature support problems. The other is TCP itself, which is incredibly flaky at web scale. As in, who knows whether a given connection will complete at all, and if it completes won't be stalled to the degree it is unusable for several seconds.

Quotes of the week

Posted Jul 19, 2012 14:56 UTC (Thu) by faassen (guest, #1676) [Link]

Kay also said some grumpy old person things about design patterns, calling them one of the most disastrous things about programming today. The argument was because we don't know anything about doing programming we shouldn't be extracting patterns, and we shouldn't give them respect.

I think that's a silly comment; how in the world are we supposed to learn more about programming without at least looking at patterns in them? We can discuss any patterns we detect on their own merits. If what they try to accomplish can be expressed better, then we can say so. If what they do can be sidestepped altogether, we can talk about that too.

Alan Kay: the web, in comparison, is a joke

Posted Jul 22, 2012 3:54 UTC (Sun) by giraffedata (subscriber, #1954) [Link]

I don't think I've seen problems with the web caused by weaknesses in TCP, but I know Kay is referring to far more basic problems than that. He's not saying it's unreliable or the features sometimes don't work. He's saying it fundamentally lacks imagination. For example, he calls Wikipedia a failure because all you can do on it is read. You go to the Logo article and you can read all about Logo, but you can't learn to use it or run Logo programs there, in spite of the fact that dozens of computers are involved in presenting the article.

A good analogy for Kay's thoughts on the web are his thoughts on PowerPoint: He hates it because all it is is a simulation of acetate slides using a computer.

Feature support doesn't seem to be an issue for Kay, because he believes features shouldn't exist. (I don't know what he means, but I can imagine a few things).

Were the web of today professionally designed (and I don't think we have to insult the designers of any web browser or protocol here; the web as a whole wasn't designed by anyone), it would be very much different. I don't know if it would find a use for TCP, but it definitely wouldn't use HTTP.

Alan Kay: the web, in comparison, is a joke

Posted Jul 24, 2012 18:45 UTC (Tue) by jwarnica (guest, #27492) [Link]

Concerning features, he expands on that a bit, I think, when discussing OSs, specifically OSs built on C. He makes the argument that because C has static early binding, you need to put all your features in when you develop the thing.

That is obviously not true in a technical sense, or in really any practical sense. However, I'm inclined to believe that anyone's view of the world is based, well, on their view of the world. Obviously there are .so's, or .ko's, with various levels of memory protection, which can be broken out of with effort. Or JVM class loading, with better protection. But its not a question of low level ASM security... more philosophy... Producing a plugin mechanism up front, which "can do anything" is possible, but invariably "can do anything" is true only for some finite value of "anything".

Quotes of the week

Posted Jul 19, 2012 14:58 UTC (Thu) by Kwi (subscriber, #59584) [Link]

The Internet was done by amateurs as well, it just has a two decade head-start over Web. And to this day, it's hardly error-free.

Witness the dozens of RFCs that fix problems in basic protocols like TCP, the painful IPv6 transition, NAT, bufferbloat, SYN attacks, Ping of Death ... and that's not even getting into the problems with core infrastructure like DNS and TLS.

And infrastructure is *easy* compared to the application layer. HTTP is a well-designed protocol compared with, say, SMTP and FTP, just to name some widely used application-layer protocols. (Of course, we all use SFTP instead of FTP, but while SFTP definitely wins on both security and features, its complexity makes HTTP seem like a triviality.)

Heck, the fact that you're reading this comment is a technological miracle.

Quotes of the week

Posted Jul 19, 2012 15:56 UTC (Thu) by sorpigal (subscriber, #36106) [Link]

I'd say it's more like the web became too popular too fast, compared to the internet. It used to be easier to fix "entrenched" architectural problems.


Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds