Officially, I was at Oracle Open World to preach about clouds. My personal goal was to listen and learn. Here’s what I’m coming away with.
Previous instalments in my OOW coverage:
Big Red — unstructured early impressions.
Day 2 — mostly keynote coverage.
Cloud @OOW — argues that there’s a substantial audience interested in clouds.
For listen-and-learn purposes, the keynote sessions were useful. The technical sessions were less so, since they were by and large aimed at very specific configurations of Oracle products; of the form “Hot New Features in PeopleSoft 9.1”.
The hallway and lunchtime conversations were still more useful. And I found that watching the #oow09 tweetstream back-channel in parallel with the action was incredibly informative.
High-Level · More of the people I talked to cared about the applications than about the middleware, and more about the middleware than the database, and more about the database than the combination of hardware and OS and so on underneath it.
Whenever the keynoters (including Larry) were focusing on the hardware or OS or I/O rates or storage capacities or replication features, there was a steady background grumble going on from people who wanted to be hearing about the apps. In particular the much-awaited new “Fusion” apps; the hunger for them was palpable.
Let me quote a much-retweeted offering from Forrester’s Paul Hamerman: “I would suggest that Oracle consider a business focused event featuring apps and processes, separate from platforms...” Yup.
Also, as a small experiment, I asked a random selection of people “Which storage products are you running on?” A half-dozen people, all from big companies whose names you’d recognize, and not one of them knew. Or, apparently, cared; I was getting these “What did he just ask?” looks.
Legacy · I also asked people about pain points, and they were all over the map, but one did stick out: dealing with legacy. The fact is that companies grow in messy non-linear unpredictable ways, and this almost inevitably leaves behind a messy non-linear inventory of business-critical infrastructure and apps.
Michael Dell in particular hammered away on this in his keynote, and based on what I heard he’s got a point; there’s nothing these people would like better than to get their houses retroactively in order.
The problem is, there’s just no easy way. If you could wave a wand and accomplish it by force of management will, we’d be there by now. I already believed that legacy migration is an area full of great big honking business opportunities, and after this conference, I think I may have been underestimating.
I was struck by the variety of technologies that got the “legacy” label pinned on them. This included COBOL, C++, WebSphere (!?), and back versions of various Oracle products.
User Interfaces · There was some negative buzz about user-interface issues. And by the standards of the public-facing social Web properties I live among, the Oracle apps are pretty humdrum. Which I’m not sure is a real problem; if an app gets out of your way and lets you manage your supply chain or your travel expenses or whatever, who cares if it’s a little drab? But there was plenty of background buzz from people whining about the look-and-feel.
There’s a Flickr set of Oracle Fusion Application screenshots; aside from lack of anti-aliasing, they look about like I’d expect from a modern business app.
Big · You can talk about legacy and apps and databases and business and technology all you want, but the #1 impression any reasonable person would have to take away from OOW is its overwhelming size. The logistics work, just barely, but the efforts to make it all happen must be titanic.
If it were a tight community of people with closely-shared common interests, that might be OK, but it’s not; people sitting at the same lunch table often ended up talking baseball, kids, and pets, after they discovered their involvements with Oracle didn’t really intersect.
What Would I Change? · My advice to the organizers of Oracle Open World would be:
Break it up. The developers, DBAs, and businesspeople just aren’t in positive-synergy mode. There are three or more manageable, focused conferences inside OOW struggling to get out.
In fact, it’s already happening; the Oracle Develop sessions were sequestered away in dingy little rooms in the Hilton, many blocks from the main venue. People were straggling into my session steadily all the way through; they’d been in the keynotes and hadn’t realized they had a long walk to a whole other venue.
In any case, if Oracle hits some of its growth targets, the conference should grow briskly and, well, I just don’t think that’s practicable.
Give the audience credit for being smart. I can’t count the number of times when, after some very straightforward demo, the presenter would say “What we just showed you was...” and then read bullet points off a slide. Gimme a break.
Cut some marketers out the loop. I found it painful, and a time-waster, to listen to a smart guy like Thomas Kurian intone those long unnatural adjective-studded sentences that had obviously been through several rounds of marketing polish (“Oh, and make sure you mention ‘best of breed’”). I think if you turned him loose to tell the audience what Oracle product development is working on and why, and what they think is important, you’d have a shorter, punchier, and immensely more compelling audience experience.
Having said all that, I was left feeling huge respect for the folk who pull this monster together and make it work pretty well.
And the people of Oracle should be proud that their products have sufficient pull to bring a gathering of this scale together.
Comment feed for ongoing:
From: Giacomo (Oct 15 2009, at 07:00)
I'd say it won't be broken up, because it's more than a conference; it's a way to project power, and a statement of intent.
[link]
From: Drew (Oct 15 2009, at 21:57)
Enterprise software always seems to be an interface disaster. Frankly those screenshots are less ugly than most business software but still look crowded and unusable. And as your post points out, most Oracle customers are probably on old legacy systems and they 9g or whatever version probably isn't nearly so polished.
In any event you have good points about the conference but it seems unlikely to happen anytime soon. Marketing's stranglehold on communications (in general I have 0 oracle specific insight) seems to be on a resurgence (now that they're wrapping their heads/hands around social media). And Giacomo is right about OOW being a statement of power.
[link]
From: JulesLt (Oct 16 2009, at 01:45)
The problem with the interface on enterprise software is that everyone pays lip service to wanting better interfaces, but no one will pay cash for them.
A lot of our clients pay us to develop bespoke web front ends for their internal users. Despite everything in the books, the number of times our developers have been able to sit down with an actual user - in the last 10 years - is zero.
In every case, there is a manager or consultant on their side representing the user.
And the number of times we've then been asked to change a screen to improve its daily usability for the office grunts using it . . . I can count on one hand. Because it will cost money - so it's easier to train users to the foibles of the system.
Usability costs - no one seems to understand that it either means building and retaining usability experience, or repeated iteration to improve things.
But the reality is that the competition is usually on how cheap can you deliver - so the pressure is on to deliver the cheapest possible system that can meet the requirement.
It's nice that their expectations of software are improving, from experiencing good consumer websites / software, but it's all users / system cost at the end of the day.
[Still, it shows you how far we've come since the mid-90s, when almost all software looked that bad]
[link]
From: Ronald Pottol (Oct 20 2009, at 20:13)
While I can see the perspective of Websphere as legacy, I mostly find that our customers are migrating from ATG Dynamo to Websphere these days.
My only comment is that Websphere is not nearly as often fully manageable from a command line (getting a root shell a box is a matter of typing "go hostname" from the jump host I am logged into, vs, perhaps firing up a vpn, then a web browser, the looking up the admin console address, port, username, and password. Gah.
[link]