This content management system is one of those which uncomfortably straddles the open source/proprietary boundary. The code is open source, but by default only comes equipped to use SQL Server as a back-end server. If you want to use it with MySQL as the back-end, you have to pay for a separate data provider. I found the project website difficult to navigate, and it was almost impossible to find installation instructions (couldn't find any on the website, and there seem to be none bundled with the download). Add to this lots of broken links, and the fact that several of the links in the Documentation page are books for sale rather than free documentation, and you've got a system which is tricky to get a handle on and install. Coupled with the fact that I'd be virtually forced to use IIS to run it (unless I can work out Mono), I'm immediately put off. A client is interested in this system, but I'm not sure I would be happy recommending it to anyone else.
This nifty piece of software could be useful. Currently supports Access and Oracle. Perhaps this is an indication of where MySQL thinks the competition is: at the two extremes of the market, the big corporate back end and the desktop. I've just watched the demo of an Access migration, and it looks really smooth. I used to use MySQL-Front to migrate Access to MySQL, when MySQL-Front was freeware and before it turned commercial. Between these two states, development halted altogether, and the product went out of step with MySQL versions. I learned an important lesson about proprietary software when this happened, leaving me stuck with a version which wouldn't work with my version of MySQL: I vowed never to use proprietary software again for anything important.
Hurrah! Google Maps UK has been launched, as looks just as good as its US counterpart. It even does directions, like the RAC. While I was playing with it, I also came across a Birmingham business directory I hadn't seen before, which might be useful.
I installed Nvu today, as I've been reformatting a document from OpenOffice.org word processor format to HTML, and it's been painful using a combination of gedit (which I hate, by the way), the Kupu editor in our Plone content management system, and the online version of HTML tidy. So far, Nvu has been great, and does what it's supposed to (WYSIWYG web page editing). I haven't taxed it much, but so far it's an improvement on manual HTML coding. I used a version bundled for Ubuntu which some helpful soul has built (it's not been properly packaged for Ubuntu apt yet), but you can probably get a version from the main Nvu site.
Barry Malzberg is a great science fiction writer. I've been reading Dwellers of the Deep, about a science fiction fan who is periodically abducted from earth, so that the aliens can try to persuade him to give him a copy of one of the magazines he collects. The dark atmosphere of helplessness is palpable, and it has the hallmark delirious rush of words I associate with his work. A nice quotation which gives a flavour of his work is this one:
Most of contemporary existence [...] is indeed a bad working draft of reality.
He really ought to be better known. Unfortunately, I think most of his 75 novels (he was very prolific in the seventies) and short story collections are out of print, though the better known ones crop up in second hand book shops quite a bit.
Is it just me, or is SourceForge actually usable these days? They've been doing hardware upgrades, which seems to be paying off. Also looking forward to their Subversion service (it pains me to say it, but CVS has always been a bit too arcane for me to grasp).
I heard about the release of the SourceLabs AMP stack through some news site, and thought I'd go and have a look at what they've done with it. I have a professional and personal interest in the idea of software stacks, and their potential as a foundation for an open source business. Anyway, the stack itself installed with no problems, and presented me with a fairly straightforward LAMP installation, no frills. Pretty good, but I won't be swapping it for XAMPP any time soon.
More interesting, though, was SourceLabs' "maven" competition, where they are giving prizes for the best forum entries about their stack. I thought I'd post something to the forum about my experiences installing the stack on Ubuntu, on the off-chance I might win. The prizes were a weekly one of a DVD of Spinal Tap and an iPod Shuffle, with a grand prize of a Marshall amp which "goes up to eleven", in homage to Spinal Tap's guitarist Nigel Tuffnel (get it? AMP, amp). So I sent in my forum posting and crossed my fingers.
About a week later, much to my surprise and delight, I got an email telling me I'd won the weekly prize! I only really wanted the DVD (Spinal Tap is close to being my all-time favourite film), as I already have a 40Gb MP3 player, but thought I could give the iPod to Nicola (my wife). I received both prizes this week. The DVD: great. The iPod Shuffle: the task of getting it working with Linux soon reared its ugly head.
Well, it turns out that there are two main programs for working with an iPod from Linux: Gtkpod and Gnupod. The former is a GUI interface, which sort of looked like it worked, but then turned out to not support the Shuffle; the latter is a command line Perl script, which does support the Shuffle, though it is a bit arduous. I built Gnupod from source, after installing the necessary Perl XML libraries via apt-get on Ubuntu. It works like a charm, but I can't see Nicola getting to grips with it. I may have to resign myself to the duty of adding new tracks when she wants them, until Gtkpod supports the Shuffle (promised in the next release). At least Gnupod works, though.
My trip to OSBC went well, and the conference was very enjoyable. I plan to write a report for my organisation's website, once I get round to it. I also have a ton of photos to put online, as I got a chance to have a look round San Francisco before the conference, and did some exploring. It's a great city. Busy this week, as I have a seminar to give this evening, and a training course on Thursday and Friday, so might have to wait until the weekend.
OpenOffice.org 2.0 has increased its dependency on Java. For example, one of the key killer features in the new version is an Access-like database called Base; this is built on top of HSQLDB, which is a Java database. The problem is: Java is not open source, which means it isn't included in pure open source distributions like Debian. So distributions have several options:
But all these strategies are less than optimal. It would be a shame to see the adoption of OpenOffice stall because of licensing, or because it cannot be installed as a standalone product (instead requiring a separate JVM download). I'm hoping that Sun will eventually see the light and open source Java: we've already seen applets fail, and J2EE floundering in the face of LAMP, and I'd hate to see OpenOffice suffer for the same reasons. For Java to reach its full potential, it needs to be included in every distribution, the same way Python and PHP are. This requires a proper open source license so it can penetrate into every open source developer's toolkit.
Microsoft is planning to shift its focus to vertical markets (solutions) rather than horizontal ones (products), according to this article. Perhaps Bill has realised he won't be able to sell many software licenses in a few years' time, as open source penetrates further into other parts of the software market (moving out of the server room and onto desktops).