Oracle Drops Sun's Commitment To Accessibility

An anonymous reader writes "What I feared has come true: after buying Sun, Oracle had a look at its accessibility group and made big cuts in it by firing the most important contributors to the Linux accessibility tools. This is a very sad day for disabled people, as it means we do not really have full-time developers any more." The coverage in OSTATIC has a few more details, including the caution: "This just shows that all too few companies are sponsoring a11y work. If one company laying off a couple of developers spells trouble for the project, then there were problems before that happened" (thanks to reader dave c-b for pointing this out).

Comments

Oracle DB

[Chris Lawrence]

Oracle has a solid core DB engine. It dates back to the seventies, but it has evolved and it's still really good. Everything built around it is pretty much crap. But people buy from Oracle for the DB engine, then get stuck buying a lot of other super-expensive, bad quality software. I love PostgreSQL, and it's getting better every day, but there's still some stuff the core Oracle engine did ten years ago you can't get anywhere else.

Re:Oracle DB

[afabbro]

• Standy databases. Yes, I know MySQL and PostgreSQL have some replication, but it's nothing like DataGuard. Do you want physical or logical? Log shipping or transactional? How about maybe you'd like to activate and test your standby database and then press a button and it's back to where it was?
• Oracle streams - a form of SQL-level replication. Master-slave, multimaster, transformational, complex business rules, etc. Nothing like it in open source.
• The whole family of Flashback: e.g., "I'd like to do a query and have the results as of the state of the database four hours ago". Or "I'd like to immediately change the database back to its state at 01:20:03am". Or "oops, I dropped a table, please bring it back instantly." Etc.
• High-performance compression that in many cases is faster than non-compression. You can encrypt it, too.
• For nearly every DB feature, Oracle has "more". It's great you have B-tree indexes - Oracle also offers bitmap and there are cases where they are really useful. It's nice that you offer hash partitioning (if you do), but Oracle can partition on a half-dozen different things. Etc.
• RAC (Real Application Clusters) - active/active (or as many "Actives" as you'd like) clusters, all instances talking to the same DB.
• Online redefinition (change your tables, views, etc. and have Oracle store everything up until you snap everything over at once - great for reducing downtimes).
• Very sophisticated introspection. By this I mean the amount of stats the DB collects on itself. There is an insane level of instrumentation and it's very easy to see where waits and delays are.
• Ability to generate and playback workloads.
• A lot of migration assistance - e.g., "here is how your database would run if you upgraded it", "here is the SQL that will not run as well if you upgrade", "here is the recommendation for fixing your PL/SQL to run better in the next version," etc.
• Query analysis is enormously better than open software (explain plans, etc.)
• Auditing is several orders of magnitude more advanced
• Star queries, OLAP, cubes, spatial, all of that.
• XML and text support are much better.
• Virtual Private Databases
• PL/SQL, Java, etc. native to the DB, as well as an entire GUI-front-end building system (Application Express)
• A fully-integrated volume/filesystem manager (ASM), cluster software, and VM, all manageable by the DB ;-) ASM is really very nice.

I'm sure I'm missing some things - those were off the top of this Oracle DBA's head. Here is a quick list of features.

I love PostgreSQL as well, and MySQL to some extent, and even SQL Server. But they're not Oracle. DB/2 is the only thing approaching its class (along with more specialized niche players like Teradata). Most of the features I mentioned above don't come into play until you're in a 24x7 high availability environment, are trying to minimize downtime, or are working at big scale.

Re:Some perspective

[WaywardGeek]

It turns out that one of the people Oracle fired is effectively the Linus Torvald of Linux accessibility. He architected it, and wrote a ton of it. It's like firing Linux, and complaining that after all, it's only one guy.

As for OpenOffice accessibility, kiss it goodbye on Linux. Without Willie or a team of several guys to replace him, it will slowly degrade in to unusablity.

I'm 100% with you on the other guys hiring Willie. My preference would be Canonical (Ubuntu), but RedHat would be a decent fit, and I could even live with Novell. Maybe they could start working off the evil taint.

Re:*Physically disabled*

[WaywardGeek]

A lot of the fear in blind-linux land isn't because Oracle fired two people, but because they fired Willie Walker. So far as I can tell from all the accessibility code I've read, Willie roughly plays the same role for open-source accessibility that Linus Torvalds plays for Linux. It's as if someone bought the company Linus works for, and said, "This guy is overpaid. Let's save some money."

I'm slowly losing my own vision, but while I can still use inaccessible software, I'm hacking like crazy in my free time to improve the things in Linux land. So, I've read a lot of code, and Willie's name is all over the place. The most important centerpiece of Linux accessibility is the Orca screen reader for the Gnome desktop. Who do you think was in charge of both Orca and Gnome accessibility? Willie, and for damned good reasons.

For guys like me who write code on Linux boxes for a living, Willie's departure from Sun is scary as hell.

Re:Capitalism at work...

[IntlHarvester]

People have been worrying about MySQL. They have been right to worry.

Its funny. With all the hubbub surrounding MySQL, hardly anyone has even bothered asking what's going to happen to OpenOffice.org.

Re:Bad title

[fm6]

I agree, the headline's wrong. But not about what got broken. When did Sun ever make a "commitment to accessibility"?

Here's what they did have: their perpetual fantasy that they could come up with a desktop that would challenge Windows. Their latest form of this fantasy was Java Desktop System, which actually has nothing to do with Java. It's just a rebranded GNOME, ported to Solaris. When I was at Sun, Sun Rays running JDS were all over the place, and JDS was heavily pushed at our customers. Though even within Sun, use of Windows or Mac PCs (usually laptops) got more and more pervasive.

JDS has to comply with federal accessibility rules, or nobody will buy it. (Nobody bought it anyway, but that's another issue.) So Sun needs GNOME to have good accessibility support. Presumably that's why Sun started contributing accessibility development. That's how all corporate contributions to OS projects happen — it isn't generosity, it's the contributor needing the product to do something it doesn't already do.

I haven't seen any announcement, but it's to be expected that Oracle will finally put an end to this expensive and futile quest for a Windows-killer. Which is why you can't find JDS anywhere on oracle.com. (The old JDS page on sun.com redirects to Oracle's Solaris page.) If Oracle doesn't need JDS, then they don't need accessibility software.

One of many Sun windmill-tilting projects that are getting the axe.

Re:*Physically disabled*

[JorDan Clock]

Because now Willie has to get a job doing something else, working many hours a week on another task that ISN'T this one. So he's left doing this work as a side-project, which I imagine doesn't leave a lot of room for serious work.