Bringing you the latest news from the Linux World.
Dedicated to keeping Linux users up-to-date, with concise
news for all interests
Sections: Main page Security Kernel Distributions Development Commerce Linux in the news Announcements Back page All in one big page
Other LWN stuff:
Archives/search
Recent features: Here is the permanent site for this page. See also: last week's LWN.
|
Leading items and editorialsAjuba Solutions has been acquired by Interwoven. This is an interesting development, and it almost certainly foreshadows events that we will see in the future. We'll look first at what's going on, then at some of the implications. Way back in the late 1980's, John Ousterhout, then at the University of California at Berkeley, put together a programming language called Tcl - the "Tool Command Language." Tcl had many of the features that have since come to be expected of scripting languages: it was interpreted, embeddable within applications, and easily extendable. With the addition of the Tk user interface library in 1990, Tcl/Tk became a widely popular tool for application development. Its mindshare may have fallen somewhat in the intervening years (and wasn't helped by Richard Stallman's famous "drive-by flaming" in 1994), but it remains popular and Tk is at the core of the GUI libraries used by Perl, Python, and other languages. After a period at Sun, Mr. Ousterhout decided to create his own company around Tcl/Tk. This company, called Scriptics, put forward a combination of open source and proprietary products, with the TclPro development environment being perhaps the flagship product on the proprietary side. Scriptics recently recast itself as Ajuba Solutions and took on a broader approach, with XML products for putting businesses on the web and such. (See John Ousterhout's Tcl history page for more on Tcl's roots). Interwoven has purchased Ajuba (announcement here) for $31 million in stock. The company has absolutely no interest in Tcl; what they wanted was Ajuba's XML expertise. So Ajuba's product line, including TclPro, will be discontinued. This is, of course, a blow to the Tcl community. The loss of the products will hurt some, but the loss of Ajuba's developers will hurt more. Ajuba was the corporate champion of Tcl/Tk, and put in a large part of the total development effort. Those developers will now be off doing proprietary XML stuff for Interwoven, and Tcl/Tk will have to do without them. Beyond the loss of developers, what are the implications for Tcl/Tk? The company has tried to answer those questions on this page about the acquisition. Among other things:
So things could be worse. There are some lessons in this series of events that are worth noting. Increasingly, free software projects have prominent corporate sponsors. Think of gcc (Red Hat), GNOME (Helix Code), Mozilla (Netscape/AOL), OpenOffice (Sun), PHP (Zend), PostgreSQL (Great Bridge), Python (BeOpen), Qt (Trolltech), and many others. This sponsorship certainly helps get the software developed and keep free software developers employed; it is thus a good thing. But the corporate world is volatile, and the tech corporate world doubly so. Fortunately, the free software community has everything it needs to cope with corporate changes - even those that are more hostile to free software than the Ajuba acquisition. Free software licenses, of course, are the first line of defense. Ajuba may no longer be in the Tcl business, but they can not take Tcl/Tk away. Software that is free will remain so. This acquisition shows, however, that it is also important to have a diverse developer base. A project that is too heavily dependent on developers at a single company will collapse if all those developers go away. Tcl is diverse enough to survive; some other projects could have a harder time. It is also important that crucial project resources live independently of the hosting company. Ajuba is working to find a new home for the Tcl development site; other companies might not bother. To this end, having sites like SourceForge around is a good thing - as long as nobody buys VA Linux Systems. It still would be nicer to see a community of SourceForges, again for the sake of diversity. Expect to see these issues come up again in the future. Linux and free software are part of the commercial world, and cannot hope to remain unaffected by it. KDE 2.0 is out. The long-awaited KDE 2.0 release is finally available. Mindful of its PR needs, the project has sent out a press release on Business Wire, complete with supporting quotes from Ransom Love, Dirk Hohndel, Gaël Duval, and others. The announcement on the KDE site is rather more satisfying, in that it skips most of the quotes and talks more about what KDE 2.0 has to offer. So, what's in KDE 2.0? At the user level there's a great many changes. Perhaps top on many peoples' lists will be KOffice and Konqueror. KOffice is the KDE office suite, which is said to be moving along nicely, though it is not yet being presented as ready for Grandma. Konqueror, instead, is the new file manager/web browser, and is said to be quite ready. It handles just about everything a web browser is supposed to do, including Java, Javascript, and SSL. There are also new window manager styles, advanced theme support, extensive internationalization capability, and, of course, the much-hyped new icons. There is a great deal of new stuff under the hood as well - much of the KDE project's effort over the last year has gone into the creation of a new advanced infrastructure. High on the list, of course, is the new KParts component system, which is claimed to be lighter-weight and easier to deal with than GNOME's CORBA-based implementation. The larger applications, such a KOffice and Konqueror, use KParts to assemble themselves out of smaller components. There is also DCOP, which allows applications to talk to each other, and KIO, supporting network-transparent I/O. The XMLGUI layer uses XML to store the details of an application's layout; it also maintains a global "style sheet" which helps to ensure consistency across the set of applications. Two years ago, critics were still saying that the free software world was not capable of producing something as complicated as a modern desktop. How much fun it is to point out to those critics that we now have two... The KDE project has raised the bar considerably with this release; congratulations to all the developers who worked to make it possible. Microsoft says penguins can mutate. Do check out this Microsoft advertisement scanned from the print version of c't magazine. It contains a set of distorted penguins, and claims that open source systems can mutate on you - so it's better to go with proprietary stuff. There are those who claim that the ad is interesting because it is the first direct Microsoft attack against free software. That, however, is not quite true - the Linux Myths page showed up on the Microsoft page just over a year ago. The ad does show that Microsoft sees a threat, though, and is looking for ways to counter it. This one is not likely to get them very far. The creator of the Word file format is not in much of a position to criticize other systems for changing - at least changes in the free software world are documented and in the open. It is an amusing ad, though, and unlikely to be the last such from that direction. Announcing...Mountain View Data. A company called Mountain View Data announced its existence this week. The first thing that catches the attention with this company is that its principals are Cliff Miller and Iris Miller, the founders of TurboLinux, along with Peter Braam, the designer behind the Coda and Intermezzo filesystems. The Millers, of course, have been easing out of their roles in TurboLinux for a while. Large venture investments have a way of pushing aside a company's founders in favor of more presentable (to investors) executives. So they are off to Mountain View, which gives them a chance to create another successful open source company. Mountain View intends to provide services around data storage needs. To that end, they have brought in Mr. Braam's Intermezzo filesystem, and Mr. Braam himself as CTO. The filesystem is being presented as an ideal high-availability solution, especially when complemented with the SnapFS filesystem (which allows taking easy snapshots of the filesystem state) and the LinuxDisk storage area network system. Mountain View is clear on its technology and personnel; what is not so clear at this point is just how the company plans to make money. The software is, after all, open source. There is a lot of talk about how corporations have increasing needs for data storage, and increasing trends toward outsourcing. Mountain View clearly plans to be involved in that outsourcing, but its services page only says "We will be offering managed storage services, early 2001." We asked Mr. Braam about what the company will sell, and were told: The business model is to manage data centers and offer storage to customers. We take care of the backups, installation, management and growth in these centers and will probably charge by the "byte" There are a few other companies operating in this area, but Mountain View is the only one explicitly working with open source software. Through the use of this software and commodity hardware, the company expects to price its offerings far below those of its competitors. It's an ambitious plan, but the company may just have the right people to pull it off. LynuxWorks files for an IPO. Just as this LWN Weekly Edition went to "press", LynuxWorks announced that it had filed for its initial public offering of stock. We've done a quick pass over the company's IPO filing, and written up our impressions as a feature article. LynuxWorks has an interesting business model in mind; it's not necessarily going to be an easy path. Inside this week's Linux Weekly News:
This Week's LWN was brought to you by:
|
October 26, 2000
|