Mon - May 3, 2004
Seaside tutorial
I spent the afternoon in the Seaside tutorial. I
won't get into it here (too. much. information.), but it was very good.
One of the biggest hurdles to using Seaside (beyond the mind blowing stack trickery) is figuring out how to use the frameworks. This tutorial was a hands on attempt to get past some of that inertia.
I don't know if I'm any more capable today than yesterday, but I got my toes wet, which is an important first step.
One of the biggest hurdles to using Seaside (beyond the mind blowing stack trickery) is figuring out how to use the frameworks. This tutorial was a hands on attempt to get past some of that inertia.
I don't know if I'm any more capable today than yesterday, but I got my toes wet, which is an important first step.
Posted at 09:47 PM Read More
Avi Bryant's Keynote on Seaside
Avi is one of the original authors of the Seaside
web-app framework, which is a novel (maybe not, but novel to me) approach to
developing web apps. One of the reasons why he was asked to speak (I presume) is
because his implementation really demonstrates some of the fundamental
advantages that a language like Smalltalk has over other, less flexible
environments. In fact, he showed us some stuff that Java (and others, Java is
just a convenient target here) is fundamentally incapable of
doing.
The talk drew a lot of inspiration from an essay of Paul Graham's about the construction of Yahoo! store that everybody should immediately read. (It talks a lot about the dangers of groupthink that seem rampant in some IT shops).
The keynote degenerated into a lot of talk about the implementation details (hey, we're all geeks here), but I think that the important message got through: we can't compete with other technologies if we play by their rules, so we should change the rules to our advantage by doing things we know they can't.
Maybe that's too much of a call to arms. Java isn't horrible, it just isn't the best tool for every job.
The talk drew a lot of inspiration from an essay of Paul Graham's about the construction of Yahoo! store that everybody should immediately read. (It talks a lot about the dangers of groupthink that seem rampant in some IT shops).
The keynote degenerated into a lot of talk about the implementation details (hey, we're all geeks here), but I think that the important message got through: we can't compete with other technologies if we play by their rules, so we should change the rules to our advantage by doing things we know they can't.
Maybe that's too much of a call to arms. Java isn't horrible, it just isn't the best tool for every job.
Posted at 09:43 PM Read More
Smalltalk & Cryptography
Start with a quick overview of cryptography
techniques (one-time pad, symmetric/asymmetric
keys)
Then, among the key-based techniques, some talk about stream vs. block-based techniques.
The presentation was supposed to be about cryptographic support within VisualWorks, but the coolest (and most effective) thing about it was that the presenter (Martin Kobetic) had gone through the trouble of assembling his slides in some software for VW that resembled book morphs in Squeak. Because his presentation was running in a live image, he was able to demonstrate some of the points he was making by running do-its.
This in itself is no big deal (who cares about seeing a string become a random byte array?), but when he was demoing the danger of reusing keys in a symmetric key system (RSA?), he ran the encoding on a clipped region of pixels within his presentation, then ran the same process (same key) on another clipped region, then overlayed the two regions, finally resulting in what was recognizable as the two clipped regions overlayed on one another.
(A xor: k) & (B xor: k) = A & B.
You had to be here, I guess. It was very effective. I think it was George Heeg who said that it was the best presentation demonstrating different cryptography techniques that he had ever seen, and I have to agree.
Then, among the key-based techniques, some talk about stream vs. block-based techniques.
The presentation was supposed to be about cryptographic support within VisualWorks, but the coolest (and most effective) thing about it was that the presenter (Martin Kobetic) had gone through the trouble of assembling his slides in some software for VW that resembled book morphs in Squeak. Because his presentation was running in a live image, he was able to demonstrate some of the points he was making by running do-its.
This in itself is no big deal (who cares about seeing a string become a random byte array?), but when he was demoing the danger of reusing keys in a symmetric key system (RSA?), he ran the encoding on a clipped region of pixels within his presentation, then ran the same process (same key) on another clipped region, then overlayed the two regions, finally resulting in what was recognizable as the two clipped regions overlayed on one another.
(A xor: k) & (B xor: k) = A & B.
You had to be here, I guess. It was very effective. I think it was George Heeg who said that it was the best presentation demonstrating different cryptography techniques that he had ever seen, and I have to agree.
Posted at 10:31 AM Read More
State Replication Protocol (SRP)
The talk started out with some talk about the
tradeoffs between a binary encoding mechanism and something more generic (XML).
The presenter (Paul Baumann) gives a little overview of his binary, generic
encoding. It's binary (efficient) but
flexible.
The code base exists in VAST but works generically in 8 dialects. One of the advantages is that you can exchange objects between dialects.
He then went into detail on the structure of the code (how to use it).
A demo using Dolphin/VAST.
The demo showed loading (into Dolphin) some simple objects as well as a complex object for which the class did not exist in Dolphin. The complex object gets loaded a a special state/structure. The complex object was then re-serialized and loaded successfully into VAST (where the class was defined).
The discussion then went in to some of the framework portability and how the same code can run in the 8 dialects. Much talk about implementation (and usage implications) followed.
The code base exists in VAST but works generically in 8 dialects. One of the advantages is that you can exchange objects between dialects.
He then went into detail on the structure of the code (how to use it).
A demo using Dolphin/VAST.
The demo showed loading (into Dolphin) some simple objects as well as a complex object for which the class did not exist in Dolphin. The complex object gets loaded a a special state/structure. The complex object was then re-serialized and loaded successfully into VAST (where the class was defined).
The discussion then went in to some of the framework portability and how the same code can run in the 8 dialects. Much talk about implementation (and usage implications) followed.
Posted at 09:23 AM Read More
Sun - May 2, 2004
Safely ensconced in my hotel room
So I've arrived in Seattle after a two day stop-over
in Whistler, B.C.
I'm on the 26th floor of the Crowne Plaza in downtown Seattle. Here's the view from my room at 11pm:
That's Interstate 5 looking Southeast. It kind of reminds me of L.A. with (of course) better air.
Being queasy about heights, I'm not about to start pressing my face up against the window, but it is sort of a neat view. Very "Lost in Translation"-esque.
I'll start blogging the conference tomorrow.
BTW, I've been reading the pragmatic programmer book on ruby. It's very readable and, with a background in Smalltalk, very familiar.
I'm trying to write a small program to data-mine the error logs we generate at work. This'll be a good introduction, I think. I'll write more as I get into it a bit.
I'm on the 26th floor of the Crowne Plaza in downtown Seattle. Here's the view from my room at 11pm:
That's Interstate 5 looking Southeast. It kind of reminds me of L.A. with (of course) better air.
Being queasy about heights, I'm not about to start pressing my face up against the window, but it is sort of a neat view. Very "Lost in Translation"-esque.
I'll start blogging the conference tomorrow.
BTW, I've been reading the pragmatic programmer book on ruby. It's very readable and, with a background in Smalltalk, very familiar.
I'm trying to write a small program to data-mine the error logs we generate at work. This'll be a good introduction, I think. I'll write more as I get into it a bit.
Posted at 11:05 PM Read More
Calendar
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
Categories
Archives
XML/RSS Feed
Statistics
Total entries in this blog:
Published On: May 03, 2004 09:47 PM
Published On: May 03, 2004 09:47 PM