Saxon repackaging

By Michael Kay on June 19, 2009 at 02:18p.m.

Saxon9.2-EE is out on beta release, so the release process has finally started. It was beginning to feel as if the light at the end of the tunnel was an optical illusion. I decided not to change the licensing. GPL is just too distasteful, and other licenses that prevent use with commercial products just seem too burdensome. So here's what I've decided to do.

There will be three Saxon editions: home, professional, and enterprise (HE, PE, EE). The enterprise edition is just a renaming of Saxon-SA, the change in name reflecting the fact that Saxon-EE now offers a lot more than just schema-awareness. For example, the streaming capabilities are now one of the major reasons people buy the product. The home edition, Saxon-HE, will continue to be open source, under the Mozilla Public License, but it will remove some of the added-value goodies that have previously been offered in Saxon-B. For example, there will be fewer Saxon-specific extensions, and less user-extensibility. The change in name, and the removal of advanced features, are both designed to reduce the amount of freeloading, that is, to persuade commercial users who are making money out of the technology to get their wallets out and feed some of that money back to Saxonica; while at the same time providing enough open-source capability for the 80% of users who only need 20% of the functionality, and who provide the potential market into which the commercial products sell.

The professional edition will be priced at around $80 for an individual license, or $1500 for a site license or distribution license. That represents two target markets: the individual user who needs a bit more capability than they can get from the home edition, and might well be paying out of their own pocket; and the corporate users who don't want to spend a lot of money and do want to keep the contractual side very simple.

Of course, this won't satisfy everyone. There may be quite a few people who stick with Saxon-B for the time being rather than "squeezing into" the smaller Saxon-HE or explaining to their boss why they need Saxon-PE (I'm well aware that it's not the money that's the issue for most people, it's the hassle of justifying it. As soon as you tell the bean-counters you want to spend some money, they ask for a business case, and that's three days out of your life that could be used for something more productive. Very few organizations measure the cost of the three days.)

And of course there are the people who will only use open source on principle. In some cases that's because their own product and business model hinges on open source, which is fair enough. In some cases it's a technical argument based on the superior quality of open source software. In some cases it's just meanness (like the people who try to ensure that it's always someone else who pays for the next round of drinks). If I make some of these people unhappy, I'm sorry (unless you're in the third group, of course). But I don't feel I owe them anything. Just because a bank has been offering free current accounts for the last ten years doesn't mean it has to keep doing it for ever. It's not as if I'm taking anything away: Saxon-B 9.1 will still be there as an option, and there was never any guarantee of a next release.

With open source there's always the risk that if you don't keep your users happy, someone will fork the code. However, Saxon is now 200K lines of code, so that's a formidable undertaking. One of the reasons I've never published the test suite is that I wanted to make forking difficult for people. I find it hard to imagine that anyone would have the stamina to make a success of it, or to win the trust of a significant number of users.

The backdrop to all this, of course, is that Saxon - certainly as an XSLT 2.0 processor - has very little competition. Basically, it's too big a product for the weekend amateur to produce, and it's too big a financial drain for the big boys to be interested (presumably this is why Intel pulled out). IBM, I see, have finally launched their XSLT 2.0 processor as an integral part of Websphere, where prices are five-figure integers: that's a realistic measure of the value people are getting, and a reasonable price point at which IBM can expect to recover their investment. I think, quite honestly, that open source did a great job in kick-starting the XML market, but is now starting to damage it. If customers expect free beer, they must expect that there will be fewer pubs. It's time the pendulum swung back a bit: some of the value that users are getting needs to be ploughed back into software development.