There’s a third part to the series of articles about building a distro that computer retailers could use to build PCs. I still don’t agree with running the project as a contest, as I’ve mentioned before, but I do think the idea in general is great. The project aims to develop not just a full distro in terms of the software that gets installed on the PC when initially setup, but also marketing materials, artwork, and apparently also software repository services. The article seems to indicate that under the concept of the Open Core of the home desktop distribution providers would now provide value added services. Currently distribution providers make self contained reasonably usable Linux systems. But the author says that the usable Linux system should be the baseline, and that disto providers should be adding something to this base. I can certainly agree with that point, but I think the arguments laid out are a bit fuzzy. I still have trouble picturing how this would unify the distribution providers. Sure, it would move the bar up, which is a good thing. It’s a step in the correct direction. But it’s not the complete solution I don’t think.
According to this article at itWorldCanada.com the lawsuit brought against IBM by SCO will not involve copyright infringement. The lawsuit will supposedly center around breech of contract, with SCO arguing that IBM introduced intellectual property belonging to SCO into the Linux kernel. I don’t know if this is really a change in position by SCO, or if this just indicates the direction against IBM only. I think they’re threatening lots of people with lawsuits, so this position used against IBM might have nothing to do with those others. I would like to hear what property exactly belongs to SCO regarding the async IO, journaling filesystem, and volume management. At the points where some of the initial work were going on for those features, I was following the Linux development lists pretty closely. And I don’t remember anyone saying that they were taking code from SCO. It seemed like lots of people hashing out ideas, writing up potential patches, and eventually coming to consensus. Perhaps SCO owns the rights to group development? Unlikely, but I wouldn’t be too surprized if they thought they did.
My power went out at home. Just a localized event, but annoying cause it happens to be localized to my area. But given the new set of toys I’m playing with I don’t have to be disconnected. I had a conversation on the Open Palm Environment WINKsite chat area using my phone, downloaded my email to my Palm, and now I’m about to post this using Vagablog. That’s pretty well connected without any power. Just another advantage of mobile computing to keep in mind.
Infrastructures.org is a great resource for anyone that needs to administer multiple Unix style machines and is looking to do so more efficiently. The principles laid down there are fantastic, and they stick to open source tools in their descriptions of how to put systems together, and even have a release of their own code called ISconf. It is hard to try to work like an infrastructure architect rather than a system administrator. But when you do the long term benefits are astounding. I’ve been places where everything was so ad-hoc that trying to setup an infrastructure around their software was futile, but even then the setup paid off in terms of keeping the base systems up to date. Any admin should at least read over the stuff. Even if you think most of it is complete bunk, there will probably be a few points you can use.
There’s an article at IBM DeveloperWorks about writing reusable command line utilities. The article focuses on five guidelines:
There’s a message on the pilot-link-devel mailing list about replacing the normal Palm Network Sync with something else. Recently Palm has been a huge pain in the ass. Just about every device comes out with some new extension to the sync protocol. Something that breaks the existing tools under Linux and other operating systems. And Palm isn’t releasing info about these changes as far as anyone can tell. So, like I also posted about the other day, the pilot-link developers are thinking about ignoring Palm’s interface for sync and writing something of their own. Something based on open protocols and standard interfaces. Something that Palm can’t change underneath us quite so often. I think there are probably a lot of people thinking in this direction. Hopefully we’ll get the critical mass required to get the project rolling. I’ve already started working on palm-xmlrpc. Something I consider a precursor to my work in the area. Hopefully this means there will be others hacking in this direction.
I have heard people talk about the danger of letting commercial software turn into the defacto standard for how to do things. I didn’t really think this was a big deal, until I tried to create a boot disk to do a BIOS update the other day. This is where the defacto standard comes in. For those who do low level system admin work on x86 based computers, DOS is the defacto standard. When a computer manufacturer wants to write an application that does something special to their hardware, they write a little DOS app. DOS is really primitive and simple and doesn’t try to stand in the way of the programs that it runs, so it’s ideal for writing specialized applications to test hardware or update firmware. I had to update the BIOS of an Intel motherboard, and the instructions said to create a boot disk and copy their utilities over to it. No problem, I don’t have a Windows system, but my girlfriend’s system should work fine. It pains me to touch the thing, but I can stand it for stints of 10 minutes or less.
The speech by Paul Saffo this evening was very interesting. There were many points he brought up that jive with the stuff that I’ve been working on. There was a lot of stuff he said that meshed well with the events I’ve seen and heard about. Here are some of the points that I found most interesting.
Just as a reminder, cause I forgot myself, Paul Saffo is speaking tonight at PARC. Should be an interesting evening, hope to see some of you there.
I created a new category for this article about a direct neural interface. The category is called Dweeb, and it’s for those things that aren’t quite mobile computing, or open source, or related to things that I’m working on, but are still just too fucking cool to ignore. I have no idea what would go into trying to build an interface like this, how well I think it would work, how marketable I think it would be… no, I just get starry eyed and whisper “oooo, I want one of those”. That’s what Dweeb is gonna be all about.
subscribe via RSS