In case anyone’s scoring at home, that’s four flights in a row for me that have suffered significant mechanical problems; today’s was a failure in the computer that controls the brakes, so as you might guess I’m pretty glad they fixed it. The silver lining is that they have all been direct flights (plus 1 for Denver), so no blown connections. But anyhow, thoughts on Day 2, my last day at the Rational conference.
Today’s keynote was delivered by none other than all around IBM and Rational guru Grady Booch, and it was a good one. From the personal history moment when he brought out a punch card program detailing the effects of bodies colliding, written when he was 12 (I was lucky to be reading at 12) to a “call” from Bill Gates (“Control-Alt-Delete, Bill”), Grady had the folks around me fairly actively engaged – and chuckling.
Of greater importance, of course, was the substance of his presentation which noted two major industry drivers, both of which I agree with:
- First, that software development has an intrinsic complexity to it that is unlikely to be entirely addressed within the foreseeable future
- Second, that the volume of code that has been produced and will be in the future is staggering (33 billion lines a year, by his estimation)
Against this tide Grady and Rational seem inclined to employ many defenses, from increasing levels of abstraction from the actual code to predefined design patterns. Rational tools, in other words. While it’s true that many hackers eschew such tools, preferring the base metal feel of products like emacs, it’s also true that as a percentage of the overall developer populuation, such highly skilled hackers are the exception rather than the rule. For that segment of the population, however, the keynote featured a lot of technologies coming out of IBM Research that really are pretty cool.
Thanks to the efforts of the Rational AR team (Amy, Cameron, and Monica), I was able to chat for a few minutes with Grady following his presentation. One of his priorities at the moment is the issue of collaborative development; i.e. geographically and potentially timezone shifted development staffs working on shared codebases. It’s actually an issue of great interest to me, because I’ve seen a fair amount of traffic in the sources I monitor in recent month concerning either frustrations with CVS, frustrations with patches & diffs included as email attachments, or frustrations with who’s working on what, when. The keynote featured a demo of “Activity Spaces,” which for those of you who know Jazz is essentially a beefed up version of that. For those of you who don’t, it’s basically a collaborative code editor that features macro integration elements like inline IM, shared tasks, etc, along with lower level niceties such as inline shared code comments. My initial reaction was that it was cool conceptually, but would run the risk of inserting a Chinese wall between the IDE and more informal collaborative mechanisms like blogs and wikis. Fortunately, the plans are to work off of existing standards such as RSS so as to leverage, rather than exclude, such resources.
I wrapped up our brief conversation with a question concerning the assertion I made here yesterday – that Rational is an organization in transition – by asking him what he saw as the most important trends for Rational going forward. To his credit, the first thing out of Grady’s mouth was simplification: simplication of the tools, simpliciation of the lifecycle, simplification of the developer experience. Simple, simple, simple. Given my bias towards simplicity in general, it should come as no surprise to hear that I was a very receptive audience for such a message. His second priority was driving home further synergies with other IBM divisions (some of the work being done around SOA with IGS is intriguing); Tivoli and Rational have created some interesting solutions together, but there’s a lot more that can be done.
Following up on another subject from yesterday, the question of what Rational’s equivalent of Gluecode might be, I unsurprisingly got little in the way of answers. I think it’s still worth watching, however, because sooner or later somebody’s going to open source some serious modelling and configuation management components (beyond Hyades), and it’ll be Rational or someone else.
As for final thoughts, I thought the show was well executed, if a bit over the top with some of the keynotes on the first day (you had to be there). Most of the grumbling I heard from other attendees had to do with the length of the walk from room to room as opposed with show content, so from a Rational perspective I’d take that as good news. Particularly for a show light on actual product news – no huge new release announcements and the like – it had a fairly good buzz going, particularly for Grady’s keynote.
As for the rest of the week, while I’m disappointed to miss a few of the sessions such as the Solutions Center demos, I can say definitively that I’m not going to miss the 100-degree-at-10-pm heat, nor the pretty much omnipresent smoking (us East coast folks aren’t used to that). I understand that I’m in the minority on this, but Vegas just is not my cup of tea – I think I’m just too much of a square. To each their own, though, and I’m sure I’ll be back sooner rather than later.