| |
 |
Saturday, 7 September 2002 |
Sell!. I'm at a small conference on the telecommunications industry. The room is lousy with very smart people, almost all Telecom Rebels of one sort or another. The general consensus: A telecommunications industry meltdown is underway. A trillion dollars in debt. No way out. It's just a matter of time — and not a lot of time — before the giants go Chapter 11. They are trying to artificially maintain the value of an outmoded network infrastructure and are in the way of delivering bandwidth to every home.
And that's the good news, relatively speaking. The telco collapse will take pension funds with it. And then real estate bubble will pop.
Drink up, my friends.
Note: I helped David Isenberg, a gen-yoo-ine telecom guru, write the Net Paradox to explain what's happening. [JOHO the Blog]
Fuck ehh?
11:59:12 PM 
|
|
Why?.  Sigh ... Something is seriously screwed up here. I posted to notes for the first time this month, and promptly got the #prefs.txt error of death. I applied the usual fix (republish entire blog, wait, try again) and it went away.
Then I posted to devlog, got the same error, tried the fix, and Radio crashed (see Figure 1). When it came back up, it was trying to upstream to radio.weblogs.com instead of pycs.net. Crash. Restart. Change community server. Republish all.
Time to make a Radio-to-bzero exporter. And a Radio template parser. [Python Community Server (development progress)]
Ahhh, so it's not just me. Thank god for that.
11:50:01 PM
|
|
Sendmail Meets Erlang.
Our software engineering team needed to create a system that moves data from a set of legacy applications with diverse properties to data repositories scattered around the network. This system had to be highly concurrent, straightforward to extend, have high performance, and be coded rapidly by a small development staff. Because of these requirements, the authors embarked upon an experiment to write this application in Erlang. This paper describes what we did, why we did it, and what we learned over the course of our development effort. It is our hope that this chronicle may be useful to others thinking about coding in Erlang for the first time and to the incumbent Erlang community to hear an outsider's perspective on this fine language.
Good to see commerical applications of functional languages. It may be interesting to compare their wish-list against what a typical theorist may want (static-typing isn't there, more code examples is). [Lambda the Ultimate]
I keep meaning to hack on erlang. I even bought the book. Hmmm.
11:46:12 PM
|
|
SchemeUnit and SchemeQL: Two Little Languages. SchemeUnit and SchemeQL: Two Little Languages. Noel Welsh, Francisco Solsona, Ian Glover.
An interesting exercise in embedding little languages (or are these really frameworks?) in Scheme.
I particularly like the discussion of SQL embedding in sections 3.1 and 3.2. It shows the advnatages that come from thinking at the programming language level. [Lambda the Ultimate]
So what is the difference between a little language and a framework? That is an interesting question
11:44:40 PM
|
|
Type Theory and Functional Programming. Type Theory and Functional Programming. Simon Thompson.
ISBN 0-201-41667-0, Addison-Wesley. 1991.
This is now out of print. I had hoped to prepare a revised version before making it available online, but sadly this hasn't happened. So, you can get the original version, warts and all, from this page.
Another great resource. Includes several nice case studies, as well as the essential mathematical foundations. [Lambda the Ultimate]
Online books. Yummy
11:43:25 PM
|
|
Solving for pattern.
Bloggers uncover the implicit and make it explici. (SOURCE:Mathemagenic)-Great idea! Bloggers build stories to make the implicit explicit.
I've been trying to articulate out what these professions have in common that could explain why weblogging has become an especially popular practice in those areas. I'm not finished thinking about it yet, but I think the commonality has to do with uncovering the implicit. Software developers patiently explain to a machine things for which humans wouldn't need an explanation. Journalists take threads from different places and build a coherent story out of them. Teachers patiently explain to students things for which trained specialists wouldn't need an explanation. Librarians gather and organize explicitly material that is only implicitly connected. Lawyers, whenever they seek to correctly interpret the intent of a law, need to uncover its spirit which is almost always implicit. All of them are not just pattern recognizers, they are also pattern explainers.
[Roland Tanglao's Weblog]
Wendell Berry has a phrase I've always liked: solving for pattern. It catches the notion that interesting answers are rarely obvious.
The other thing that these professions share is that they are all people who think for a living. As such, they've developed a healthy respect for the time and effort it takes to develop a decent thought. This may be less evident to bloggers than it might otherwise seem. If you write or even read blogs regularly, you're accustomed to thinking and may not recognize how rare a skill it is (it is a skill and, therefore, something that can be developed). [McGee's Musings]
Pattern recognition is an area where computers historically have done very poorly. With the use of fuzzy neural nets computers aren't so bad anymore. A question in my mind is if it is possible to feed text into such a FuzzyNN and get useful stuff out. There are a myriad of ways of configuring such a beast, but are there ways that generate useful output on current desktop computers?
11:30:06 PM
|
|
Bayes and so forth. There has recently been a lot of fuzz about spam filtering with naive Bayes networks. Naturally the theme got caught in the python-dev too. Several Python developers have played around with this scheme and result can be found in Python CVS nondist/sandbox link. Tim Peters now wants move on to other things and gave final thoughts on the subject link. For some reason you are somehow inclined to make conclusions about different algorithms before you have any verification. Therefore Tim advices that:
The second most important thing is to run a fat test all the way to the end before concluding anything. A subtler point is that you should never keep a change that doesn't *prove* itself a winner: neutral changes bloat your code with proven irrelevancies that will come back to make your life harder later, in part because they'll randomly interfere with future changes in ways that make it harder to recognize a significant change when you stumble into one.
and
there's-no-substitute-for-experiment-except-drugs-ly y'rs - tim
So, in general, if you are working on Bayesian spam filtering, I would suggest you to take look at python-dev link archives, too. [Python owns us]
Cool
9:07:05 PM
|
|
The Networking Computing blog is zooming (subscribed!). If you are publishing an IT industry e-zine (or particularly if you aren't) and not publishing a Radio weblog you are missing out. [John Robb's Radio Weblog]
Argh more blogs
9:00:18 PM
|
|
Simpleface.org is now a weblog-wiki. I finally got a chance to add the weblog front end to the Simpleface.org website. It wasn't very hard to do at all, I just ripped up the MiniBlog code and did some welding at the seams. Since they both use the exact same table structure already, there was only concept hurdles to overcome.
I'm going to use the Simpleface.org site as a testbed for a few days, then I'll integrate the code into Simpleweb for those who want... I guess I could do it now, but I want to add comments in first (which aren't done just yet). And it's already 3:18 a.m. (doh!)
Now, I didn't come up with the weblog-wiki concept, but I've decided to name it: I'm going to call it a Bliki or a Wiblog. What do you think?
More thoughts: The Bliki concept extends weblogging in a way that I didn't realize until I started testing it. Each post is a more permanent webpage. The basic difference is that where a weblog refers to posts by date, a wiki refers to the page by name, this gives each individual page a lot more weight, in my opinion. Therefore, I think what happens is the posts become heavier - leaning towards more or less one post per day - rather than this weblog for example, where each post is a separate record. This means that instead of adding posts for every little thing, it'll probably be better to add only one or two posts per day, and continually edit those pages with new info. It makes permalinks less finite, but it'll make the long term pages more accessible.
That's one of the things that's sad about blogging. Much of the info you write disappears when it goes off your front page. I've gone back and read some previous posts of other bloggers, but generally I don't unless they themselves direct me there. I start reading from today on... All those ideas and thoughts from last month are like ether. A Wiblog corrects this by making the posts mean a bit more - they're accessible and easy to create as a weblog, but will have the staying power of a real web page. Or... well, that's my first impressions.
Off to bed.
-Russ [Russell Beattie Notebook]
I have to admit I have dreamed about the capability to automatically triage the flow of tidbits flowing through my blog into something Wiki-ish.
8:20:51 PM
|
|
Introducing Nonblocking Sockets. Nonblocking sockets, the new I/O API in J2SE 1.4, allow I/O operations on a channel without blocking the processes using it. This will allow for asynchronous high-performance read/write operations that will totally alter the techniques for developing socked-based applications. [O'Reilly Network Articles]
If I ever get back to hacking network code directly this could be useful. Assuming that I can get to develop code for j2se installs later than 1.2 :P
8:13:36 PM
|
|
Microsoft Execs Reap Bitter Harvest. Bill Gates has never had a reputation for government cooperation. But the renegade
attitude that led to his riches in the technology world ultimately could lead to the
destruction of his software empire. Indeed, fallout from the various antitrust lawsuits
-- either direct or indirect -- is becoming more and more apparent. [osOpinion]
Having watched more than one dominant player in this industry be turned to dust, it will be interesting to see how Microsoft copes. They can either continue trying to snuggle up to Hollywood and get crushed by customer indifference, or they can start shipping complete developer sdk's on all new WinXP machines and watch the uptake sky rocket. Microsofts golden hen is independent coders who bring change and vitality to the platform. They have completely destroyed the market for independents by repeated suffocation.
Is it only me that finds the fact that almost all software for windows is release 5 or above scary? It's dying of senility, or maybe boredom.
8:07:24 PM
|
|
My guilty secret. Just been talking to Scott Johnson and I've realised I have a guilty secret.
I admit it.
I am a 30-something and I enjoy watching Buffy.
There, I feel better now.
Actually the thing that really irks me is that I didn't get into it sooner. I missed the humour the first time round and ignored the show. But thanks to Sky's policy of endlessly re-running everything I don't have to miss out. [Curiouser and curiouser!]
I have to admit the major reason I want to buy a DVD player is so that I can buy all seven seasons of ST:TNG and watch them in order. I suspect that I will also be buying all the Buffy and Angel series too. I love crap TV - I just hate having to stay up to watch it, and putting up with adverts for fone sex lines. Heh.
7:52:32 PM
|
|
New Languages on the JVM.
Just had another thought. We've got Jython, Rhino and Jacl (Python, JavaScript and Tcl? for the JVM). So why can't the JVM be the non-proprietary common runtime platform for scripting languages to share? I wonder what extensions to the JVM would be required to really support all languages and all platforms. [James Strachan's Radio Weblog]
Yes, exactly. I can't understand why there's this apparent bias against compiling other languages into Java bytecode. I'm pretty ignorant of the details of the JVM, but my understanding is that it's fundamentally the same as the .NET CLR; a stack based register machine. IIRC, both lack support for some fundamental constructs that functional languages find convenient, but these can be emulated, with a performance penalty. So as it is, Java bytecode is OK as an IL. The question is, what kind of support is there for bytecode generation? It seems like a lot of the langagues (compiled ones, at least) generate Java code that has to be compiled and this is an obstacle. Code makes a lousy IL, though JSP has done well with it, but JSP allows embedding literal Java, and really is a macro processor, not a language. To me, one of the huge strengths of the .NET FCL is the support for dynamic code generation (CodeDom), compilation (ICodeCompiler) and most importantly for this topic, IL generation (System.Reflect.Emit).
It occurs to me that there's possibly another problem here. While objects are baked into both VMs, I'm not sure if .NET requires all code to live inside an object. I'll have to look into that. Anyway, it seems like there's a problem with mapping non-OO languages into an OO VM. There's probably a way to just map everything into a single global object though, the equivalent Java code wouldn't be pretty, but Java code isn't the point. [Gordon Weakliem's Radio Weblog]
IIRC Microsoft Research and several others were asked to push various languages on top of CLR. Out of that research the requirement for tail recursion capable, which was included in CLR. So, CLR is more cross language capable than JVM (which is tail recursion incapable AFAIK).
The large chunk of learning that both Java and D Flat failed to learn from LISP and Scheme that the real thing required to do multi-language hacking is the ability to effortlessly bootstrap small languages via powerful macro-system. I'm a great fan of things like XDoclet, but the kicker is that to do something like XDoclet in scheme is orders of magnitude easier. Sad but True.
7:47:08 PM
|
|
KEN LAYNE:. "Ain't no regime in the world strong enough to deal with an army of pissed-off teen-aged blogger girls." [evhead]
It will be fun watching the fur fly. Errr. Maybe that was a bad choice of words ....
7:38:50 PM
|
|
Think Outside the Browser. Unfortunately the goals are framed in terms of creating web sites rather than finding a way to deliver a service that may include the use of the web.
Even if there weren't performance problems on today's Internet the local PC is where the high performance interactions take place. And the PC should act as your agent in dealing with the remote services rather than burdening you with the quirks of the raw Internet.
Bob Frankston, "Trapped by the Web!," www.satn.org, 2 Sep 2002 via Tomalak's Realm [ideas]
This time is coming. But, it would be nice not to lose the (almost) cross platform rapid development advantage of hacking html and cgi scripts.
7:36:01 PM
|
|
Everything's explained in SOP. In many organizations, the design process still follows a classical "military" model. A small number of "officers" do the "noble" task of analysis and design, plan the overall work load, and assign implementations and test tasks to the "simple soldiers." This way of working has several drawbacks....But today, even the military no longer works like this, at least in the British SAS or American SEALS and probably other elite military forces as well. ... Most interesting to us is the account of the... [short notes]
I wish I had a reference for this when I was doing a recent assignment. I was comparing and contrasting three SDLC's - Waterfall, Spiral and XP. I must say, speaking as a coder, XP rocks. I can see, however, that it gives those used to being able to map out a Gantt chart of an entire project absolute heart ache...
7:31:20 PM
|
|
Plastic Optical Fibre: Cheap and Bendy [Slashdot]
Sadest part about this? In Oz, after the abject failure of national cable TV roll-outs there will be no VC money willing to replace last mile copper with this. And thus, we will be leap-frogged by the new nations on the planet that haven't yet rolled out infrastructure. Think Asia. Think India. Ouch.
1:35:51 PM
|
|
GASOLINE ENGINES CAN BE VIRTUALLY POLLUTION-FREE according to an engineering study done by the University of California, Riverside. [InstaPundit]
Gee, I wonder if you can make polution free methanol/ethanol motors?
1:33:22 PM
|
|
Generics
Yesterday, Microsoft Research released a beta version of Gyro, which is their implementation of generics for Rotor.
I downloaded it this morning and played with it, and I must say, it's great! This is going to be the killer feature for .NET (once the commercial platform gets it). Unlike the platform that starts with a J, this is actually done right. The Java generics proposal that was floated for 1.4 and then withdrawn was an ugly (nay, disgusting) hack. Gyro is beautiful, I have no other word for it.
I can't wait for this to appear in the commercial platform! [IK.VM.NET Weblog]
Welp, as far as I am aware Generics are slated for release in 1.5. It will be interesting to see how Java and D Flat compare.
1:27:17 PM
|
|
java.blogs. Gonna wear that new icon proudly. Check out Mike's list of Java blogs.... [paradox1x]
Cool. I wonder if there are any blogs in there I am not addicted subscribed to.
1:14:35 PM
|
|
Zen. If a man says something, and there's no woman to hear, is he still wrong? [The Peanut Gallery]
Yes.
1:07:32 PM
|
|
Canadian Senate committee recommends legalizing pot. The Canadian Senate Special Committee on Illegal Drugs has just released a report (PDF summary) that recommends legalizing marijuana, and wiping the records of those convicted of posession. According to committee chair Senator Pierre Claude Nolin, "Prohibition is a cop-out." Canadian news coverage from CBC, CTV, and The Globe and Mail. [kuro5hin.org]
I can see the future now, America invades Canada to prevent "Dope Head Terrorists" from underming the trans-global war on drugs ...
1:00:50 PM
|
|
All electronic gaming outlawed in Greece.. In Greece, law 3037 (.pdf in Greek) has come into effect on the 30/7/2002. This law bans all wholly or partially electronic gaming, both in public and in the home. [kuro5hin.org]
I can't quite figure if this is insane or inane. They can kiss their IT economy good bye however.
12:58:18 PM
|
|
Interesting Java Analyzer. DoctorJ will go through your Java code and check for all sorts of common mistakes, including spelling mistakes in your javadocs and unnecessary import statements. It looks like something that could be really useful [weblog.masukomi.org]
INteresting
12:56:27 PM
|
|
Going Up?. Perl 5.8.0 brought stable threading to Perl - but what does it mean and how can we use it? Get a lift with Sam Tregar as he creates a multi-threaded simulation. [Perl.com Perl.com]
I will probably need to look into this at some point. If only for fun.
12:48:00 PM
|
|
Installing Nagios. Nagios is a system and network monitoring application. It watches hosts and services that you specify, alerting you when things go bad and when they get better. [O'Reilly Network PHP DevCenter]
Anyone want to offer opinions as to how this compares with say, BigBrother?
12:22:22 PM
|
|
Consume this
Take a look at what Hollywood's working on to replace the Net.
I wonder if this will go the way of DAT tapes & MiniCDs? Actually Hollywood is going to have a difficult time displacing DVDs. They are the entrenched tech now, and within a year or two start expecting to see chinese made players that can play DVDs thanks to th power of DeCSS....
12:16:37 PM
|
|
© Copyright 2002 Brett Morgan.
|
|
|
|
|
