On the 6th of May, the Apache community took some major steps to further “open up” Java. On this date, a proposal was submitted to create Project Harmony. The proposal requested sponsorship/support for a project to build an open source J2SE implmentation. The proposed project would:

1. Create a Compatible, independent implementation of J2SE 5 under the Apache License v2.
2. Create a community-developed modular runtime (VM and class library) architecture to allow independent implementations to share runtime components, and allow independent innovation in runtime components.

This doesn’t sound like much. After all, there have been open source JVM’s (JSR’s) for quite some time. But this is a big deal. Under the terms of previous Java licenses, it was not possible to build a completely open J2SE environment. But the Sun Java 5 license no longer precludes such a development. Consequently, a proposal was submitted to create a completely, open source J2SE. But some of the really interesting pieces of the puzzle weren’t announced with the proposal. They include the following:

• Apache is not looking to supplant Kafe or Classpath with Harmony. Rather, I think the project team membership is inidcative of widespread support for a community-developed J2SE implementation – including comprehensive class libraries! If anything, the community might well desire the integration of these currently divergent efforts. I know that I do!
• IBM has an interesting “connection” to this project. As some of you may remember, IBM recently acquired Gluecode. So far, so good. And one of the principals at Gluecode is Geir Magnusson. Geir has been the project leader for the Apache Jakarta project since early 2004. He was also the lead noted on the Apache Harmony proposal. And Geir is now an excited IBM employee. While Gluecode was not purchased just to promote Harmony, there are very strong ties between Gluecode and Apache. Indeed, Gluecode has been one of the premier sponsors of the Apache Geronimo project (which would provide J2EE/EJB capabilities in an open source container).

I believe that with the acquisition of Gluecode, IBM is making a clear statement that its Java strategy will include a commodity investment (in Jakarta, Harmony, Geronimo and other Apache efforts) as well as its “boutique” WebSphere line. While the final outcome of this market differentiation is not yet clear, it is clear that IBM wants to put some pressure on the JBoss incursions into WebSphere territory. As for me, I will watch whether or not IBM also decides to invest in the Apache Portals projects.
Finally, I think it’s fascinating to note the comments from some of Sun’s Java luminaries. Graham Hamilton was positively ebullient about Harmony. Indeed, he wishes the Apache Foundation success on its efforts. And Simon Phipps warmly welcomed the Harmony project into existence. I wonder what Mr. Gosling thinks about these developments.
-CyclingRoo-

I’ve been trying to find a good way of describing Greasemonkey to friends and colleagues. It’s been a little dicey. Some folke get it immediately. Other folks are much more visually centered. So in my efforts to find a quick and easy way of explaining Greasemonkey, I ran across a screencast of it.
Jon Udell explains screencasting this way:

Screencasting is a medium that can deliver order-of-magnitude improvements over conventional documentation and training. If people aren’t “getting it,” maybe it’s time to stop blaming them and start telling stories they can’t forget or misunderstand.

Put more simply, its video and audio of computer screens, transitions, and effects. Hence, it is perfect as a means for remote training of targeted content.
And for those who do want a wide assortment of Greasemonkey info, check out the Dive Into Greasemonkey site.
-CyclingRoo-

Bits of code and interesting tidbits…
1. The official Firefox 1.0.4 release is now available. It can be downloaded here. The MoFo team rocks. When confronted with the security challenge, they accepted responsibility for code errors. They had a patch in test within 48 hours and the final bits were QA tested for quite a while. This proves one thing. The open source model doesn’t create perfect code. But it does address code errors more directly and efficiently!
2. I have previously noted that Greasemonkey is a very cool tool. It puts browser and content control more fully into the user’s hands. Well, folks are taking this one step further. Asa Dotzler is now talking about a tool called Platypus. In Asa’s words,

I can remove, hide, and move elements around on a page. I can insert arbitrary HTML. I can change style rules. I can modify URLs on a page. Then I can save all those changes as a greasemonkey script that will run when next I load the page restoring it to my “improved” state.

This is truly cool. I’ve downloaded the tool and will report results as they emerge.
-CyclingRoo-

I just saw an InfoWorld article that has me truly excited! Based upon this article, Blogger will be getting new photo/image management capabilities. That will be great. While Hello is OK, it is very cumbersome. And InfoWorld is noting that this integration may include GMail integration. This would be immensely welcome – assuming that you have GMail.
So here’s my question to Google: will GMail invites be sent to all Blogger accounts? I don’t need one as I already have a GMail account. But what about the other Blogger users?
-CyclingRoo-

SVG (Scalable Vector Graphics) has been a web standard for almost two years. Adobe has had a plugin (at least, for Windows) for quite some time now. But there are no production browsers that provide native support for SVG.
But that’s about to change. Both Firefox and Opera are preparing new releases that will incorporate native SVG support. MoFo (Mozilla Foundation) is on record stating that native support will be part of Firefox 1.1. But on Tuesday (April 26th), Chase Phillips noted that native SVG support is now part of the nightly builds for Firefox.
So I grabbed the nightlies and voila! My browser now supports native SVG. Now I can see the OpenNMS maps w/o the need for a plugin. And there is so much more to be found in the SVG universe. So run over to your favorite mirror and get an SVG-enabled build. You’ll need to activate SVG (by setting svg.enabled to true in about:config). Then go and check out this link. But even better than the fixed samples is this SVG-enabled Tetris game.
Great stuff!
-CyclingRoo-

This post is long overdue. But that’s my fault for not wanting to post an announcement before I had a chance to test and deploy the new release. When last I posted about OpenNMS, I had just upgraded our infrastructure to v1.1.5 (the release candidate for v1.2). The upgrade to v1.1.5 was difficult – but worthwhile. I had to struggle through data validation errors and other migration challenges. But after about ten hours, I was able to start the production server with the v1.1.5 code.
Well, I didn’t want to take that kind of an outage again. After all, we have ~350 nodes and ~750 interfaces. That’s not a lot. And since v1.1.5 was the release candidate for the “soon-to-be” stable branch, I thought I’d wait until the code stabilized – and the unexpected bugs were swatted. And, true enough, there were two quick release after v1.2.0 was originally released. So I waited until yesterday – and dove in.
I preped the box by downloading/moving the RPM’s and backing up all the customized configuration files and html pages. So at lunch, I brought the system down and started the upgrade. [BTW, I love daytime outages. It sure beats the Saturday night + Sunday morning outage windows in most businesses.]
Unlike the v1.1.5 upgrade, this one went flawlessly. The software installed w/o any problems. And the installation script validated all the necessary table dependenices – most of which had already been ironed out with the v1.1.5 release. After the software install and data prep, all I needed to do was apply the customized html edits and all replacement pages. After a quick daemon restart, the system came up flawlessly. Props to Tarus, DJ and the crew. The new stable branch rocks! It is fast and full of cool new features.
Total install time: 15 minutes
-CyclingRoo-

Given Intel’s recent announcement of WiMax chipsets, Robert Cringely has decided to opine on WiMax and its market potential. I don’t want to outline all the advantages and disadvantages of WiMax. You can read Cringely’s article and/or Google the subject. Nevertheless, the recent announcements make it a subject worth a few lines of commentary. So here are my idle thoughts on the subject.
WiMax has two roles: broadband extension and alternative broadband access.
In its first role, WiMax can be used as a broadband extension technology. Using unlicened spectrum at 5.8GHz, you could extend your broadband access to your neighborhood, but not much further. Indeed, this might be an interesting way of raising money within a homeowners’ association – assuming you have sufficient bandwidth on the back-end. But unlicensed spectrum holds little promise beyond extension of existing access. The reason is simple: to boost a signal to go any substantive distance, you need to be much “louder” than other things in the area. And the FCC really tries to ensure that people are not abusing the EM spectrum.
Nevertheless, unlicensed WiMax will provide some interesting extension possibilities – however localized they may be. I expect to see some entrepeneurial neighborhood networks. But I also expect to see city market deployments and/or town center kinds of deployments – at least, in areas where WiFi mesh networks have not already penetrated.
In its second role, WiMax can be used as an alternative to DSL and cable. Using licensed spectrum, WiMax can connect end points at up to 20 miles distance (and maybe even further). So it would be feasible to extend broadband out into more rural areas by dropping a DS3 in a central location and then connecting it to hubs that are wirelessly extended beyond traditional CO distances (or cable head ends for that matter). In short, WiMax holds the promise of extending wireless broadband to areas where wired broadband might not be able to economically reach.
But this will require licensed spectrum. And spectrum costs real money. And who has the money: telcos and cable providers. And they might have disincentives to such investments. Namely, they would be investing in a competitive technology that might threaten their existing installed (and depreciated capital) base. So telcos and cable service providers won’t have immediate incentives in markets where they already have products. Fortunately, there is enough uncovered territory for the behemoths to fight over.
At the same time, you might see new competition in the lesser-established markets – assuming the economics make sense. A new competitior would only need to hook into existing metro area networks or loops. And what geopolitical features do most fiber runs follow? Usually, they follow highways and or railways. And since most people live around transportation channels (like highways or railways), these seem like the right places to connect into the fiber infrastructure. And who controls rights to highways and railways? Usually, it’s the government (either municipal, state or federal).
So where is this long train of thought leading? OK. I’ll get to the chase. I believe that the muni-wireless discussion will spur many communities to adopt a WiMax distribution model centered around access “hubs” positioned along existing transportation channels. On federal and state highways, I would look to the rest stops as the most likely place to house such infrastructure. And in more isolated communities, I would look to train stations or switch yards as likely targets for the WiMax distribution infrastructure.
In any case, the prize will go to the groups that get connected to the distribution channel first. So the early bets will be on telcos and cable suppliers. After all, they do have the capital advantage. But I wouldn’t be surprised if a number of nimble providers pop up in each market. The trick will be to coordinate the eventual consolidation of all of these smaller providers.
-CyclingRoo-

Given Intel’s recent announcement of WiMax chipsets, Robert Cringely has decided to opine on WiMax and its market potential. I don’t want to outline all the advantages and disadvantages of WiMax. You can read Cringely’s article and/or Google the subject. Nevertheless, the recent announcements make it a subject worth a few lines of commentary. So here are my idle thoughts on the subject.

WiMax has two roles: broadband extension and alternative broadband access.

In its first role, WiMax can be used as a broadband extension technology. Using unlicened spectrum at 5.8GHz, you could extend your broadband access to your neighborhood, but not much further. Indeed, this might be an interesting way of raising money within a homeowners’ association – assuming you have sufficient bandwidth on the back-end. But unlicensed spectrum holds little promise beyond extension of existing access. The reason is simple: to boost a signal to go any substantive distance, you need to be much “louder” than other things in the area. And the FCC really tries to ensure that people are not abusing the EM spectrum.

Nevertheless, unlicensed WiMax will provide some interesting extension possibilities – however localized they may be. I expect to see some entrepeneurial neighborhood networks. But I also expect to see city market deployments and/or town center kinds of deployments – at least, in areas where WiFi mesh networks have not already penetrated.

In its second role, WiMax can be used as an alternative to DSL and cable. Using licensed spectrum, WiMax can connect end points at up to 20 miles distance (and maybe even further). So it would be feasible to extend broadband out into more rural areas by dropping a DS3 in a central location and then connecting it to hubs that are wirelessly extended beyond traditional CO distances (or cable head ends for that matter). In short, WiMax holds the promise of extending wireless broadband to areas where wired broadband might not be able to economically reach.

But this will require licensed spectrum. And spectrum costs real money. And who has the money: telcos and cable providers. And they might have disincentives to such investments. Namely, they would be investing in a competitive technology that might threaten their existing installed (and depreciated capital) base. So telcos and cable service providers won’t have immediate incentives in markets where they already have products. Fortunately, there is enough uncovered territory for the behemoths to fight over.

At the same time, you might see new competition in the lesser-established markets – assuming the economics make sense. A new competitior would only need to hook into existing metro area networks or loops. And what geopolitical features do most fiber runs follow? Usually, they follow highways and or railways. And since most people live around transportation channels (like highways or railways), these seem like the right places to connect into the fiber infrastructure. And who controls rights to highways and railways? Usually, it’s the government (either municipal, state or federal).

So where is this long train of thought leading? OK. I’ll get to the chase. I believe that the muni-wireless discussion will spur many communities to adopt a WiMax distribution model centered around access “hubs” positioned along existing transportation channels. On federal and state highways, I would look to the rest stops as the most likely place to house such infrastructure. And in more isolated communities, I would look to train stations or switch yards as likely targets for the WiMax distribution infrastructure.

In any case, the prize will go to the groups that get connected to the distribution channel first. So the early bets will be on telcos and cable suppliers. After all, they do have the capital advantage. But I wouldn’t be surprised if a number of nimble providers pop up in each market. The trick will be to coordinate the eventual consolidation of all of these smaller providers.

-CyclingRoo-