| |
 |
Thursday, 19 September 2002 |
Mozilla Rising. Salon.com has a good article about Mozilla, Netscape, and that crappy browser known as IE that hasn't improved noticeably since they ran Netscape out of town.
It also has a mention of Multizilla. If you haven't tried Mozilla since it went 1.0, you owe it to yourself. After you've got it running, go get Multizilla and enjoy a whole new world of browsing. [Jason Carreira]
This got me excited enough to install a mozilla nightly I had on a recent backup cd from my work lappy, but according to the MultiZilla home page:
Mozilla recently changed the skin version number from 1.0 to 1.2. This breaks most of the current mozilla add-ons, but we have made our latest, but still experimental, MultiZilla v1.1.22 release available for download. This version has been confirmed to work with mozilla 2002091309 on Windows/Linux but should *not* be used on a Mac and mozilla builds prior to September the 13th!
I want a way to download nightly install deltas. Downloading the entirety of the downloader over a 56k modem takes way too long.
11:26:11 PM 
|
|
Distributed objects and collaborative content filtering. Rickard Said:
Wouldn't this be nice: with the AOP framework I've described clients can use the object model directly. The server uses, basically, a persistent hashtable to store them. Now, what if that hashtable was a JavaSpace? Then the objects would be "distributed" and accessible everywhere, but you wouldn't have to know exactly where they are, just that they're in the JavaSpace. Then, what if that JavaSpace was backed by a JXTA peer2peer network? Then you could have islands of JavaSpaces that can talk to each other over considerable distances and have those networks created through the JXTA framework. Then, what if (some) objects in this network of JavaSpaces could migrate to whereever it is being used "the most" at the moment?
They're everywhere : Random Thoughts 9/12/02
What if they were built on JavaGroups? I must admit, I haven't read anything about JXTA, and I've only skimmed the ideas behind JavaGroups, but I really like the protocol stack idea in JavaGroups, and the stuff they're doing with clustering JBoss is VERY cool. This could be an interesting basis for the collaborative filtering / ranking part of the RSS feed syndication. [Jason Carreira]
I think my only bitch with JXTA was that when I looked at it they were desperatly trying to remain jdk1.1 compliant for applet use. There are more interoperable messaging systems out there that may be worth thinking about. But, from a first glance JavaGroups looks interesting.
11:11:56 PM
|
|
Thinking about a distributed rating system. One of the things that annoys me about Radio Userland is how it shows the articles grouped by source. I want to see all of the newest, most relevant articles at the top. I'm thinking that the distributed rating system should assign each article a score based on a few parameters:
- Your rating of this article
- Age of article (inversely proportional)
- Your rating of this source
- The rating of the people you've added to your trusted group
- The rating of the people in their trusted groups (lower influence)
- Number of articles referencing the article - not sure how this could be done, it really needs something more complex than just RSS, but it would be awesome to add this kind of "google" link rating
If you add filters on top of this, to allow you to only get articles with certain keywords, etc., and I think you've got a powerful means of sorting information to see what's the most relevant. [Jason Carreira]
Funnily enough I have been thinking similar thoughts, except with a fuzzy logic flavour. I'd think it hyper cool if said ratings and rated articles could be exportable such that people interested in trialing odd shaped AI bots could do so. Not that I know such people. Honest.
11:07:21 PM
|
|
Spam blocking coming to Mozilla Mail. Browsing the CVS checkins for Mozilla, I noticed a lot of work being done on a spam blocking feature for Mozilla. After a little digging I found the bugzilla bug associated with it, 169638. It seems like this feature will be based around a "scoring" system that queries email to [Blogzilla - a blog about Mozilla]
This could have me using Mozilla as a mail client quick smart.
10:56:48 PM
|
|
Hildebrandt: Relationship Intelligence for Law Firms. Joy beats me to the punch: Today, Hildebrandt's TechGroup and my company (Interface Software) published a white paper about the importance of Relationship Intelligence in law firms. Joy's comments:
Relationship Intelligence and Law Firms. The Hildebrandt TechGroup, a division of law office management consultancy, Hildebrandt International, published a white paper entitled: Relationship Intelligence in a Competitive Market.
With information provided by Interface Software, the white paper is an excellent law firm CRM primer. The white paper also describes "The Partnership's Dilemma:"
"The nature of the dilemma is that profit squeezes are hindering investments in a new business model that incorporates new programs such as practice management, knowledge management and client relationship management. The new programs require new staff and new systems—just at a time when firms are compelled to spend more on these line items just to maintain their current capabilities. At the same time, many firms are also experiencing downward pricing pressure . . . " [excited utterances] [tins ::: Rick Klau's weblog]
Maybe lawyers need to look into new business models. They could approach law akin to software development, where they build up knowledge relating to some area of the law into a knowledge cluster in the heads of various staff members (via a KM of course), and then charge clients based both on some percentage of hours spent, as well as some portion of the cost of the knowledge base they have formed. The percentage of the hours uncharged goes into the cost base of the data store. Then you wind up needing a way of rating how useful the knowledge base is. I can see lots of interesting approaches to abuse costing models.
The most important thing is that law firms, like software houses, need to recognise that people are not interchangeable, and their worth is related to how well they take information in, and give knowledge out. This is not a simple problem.
10:54:08 PM
|
|
Well memories of a good night are slowly fading. Even if I did turn up two and a half hours late. Interesting discussion of game theory which somehow lead to communism vs socialism vs fascism, in the context of comparing sweden and singapore. It had something to do with statistics I think. Which is why I have another book on my wish list. I think the best point during the course of this conversation was Charles' call for more Swedish Pr0n Stars. Anyway, lessons learned:
- Don't get caught in strategic meeting at 6pm. It will not finish until 9pm.
- Don't go looking for Indian resteraunts in Newtown at 12:30am on a wednesday night.
- Don't ask scotsmen about what goes into Haggis.
All in all, can't wait for the next installment.
10:44:16 PM
|
|
How many project managers that you know would accept a project
like the following:
"We are merging two databases into one database, and we need you to
confirm that the 150 odd applications from the database to
disappear will work on the new database after cut over. The
cut over is in two weeks, the applications have no test
cases, documentation, or data dictionaries, and the people who
wrote the applications have mostly left, have been re-assigned, or
fired."
Well, I know one that did. According to a class mate who works for one of those accounting/auditing firm IT consulting spinoffs this is regular. Expected even. How the hell do large companies stay alive with this level of incompetence?
10:29:06 PM
|
|
You'll have a job. This excellent post presents 12 reasons why a good programmer should not worry about employment. It comes from the "I couldn't say it better myself" department. I remember discussion with other programmers where they claimed that because VB and other such tools make programming easier, there will be less work and less demand for programmers. To which I say: bull. Those tools make some programming easier but at the same time a lot of programming becomes harder (because our software gets bigger), there are tons of new software that begs to be written and every year we tend to come up with a new, uncharted territory waiting to be filled with new software (web servers anyone?, IM systems, weblog software, native XML databases, web search engines - the list goes on and on). There'll be plenty of work for good people. [Krzysztof Kowalczyk's Weblog] [Ron Lusk's Radio Weblog]
I remember telling my dad when I was about 10 that by the time I grew up there would be no programmers, because computers were being used to write programs. I was worried, because my dad earned a crust as a IBM mid-range programmer. This was about the time of the Japanese playing with Prolog and the like.
These days I iz a programmer. Weird but true.
10:12:39 PM
|
|
© Copyright 2002 Brett Morgan.
|
|
|
|
|
