alanc @ sun.com

X Dev Conference so far...

I'm in Boston this week for the X Developer's Conference. It's been a very interesting time - good to meet many people, some that I've talked to before on the phone or e-mail, others that I haven't.

I won't summarize the talks - Federico Mena-Quintero is doing that in his blog. The environment isn't quite what I expected, but is working well. The meetings are broadcast via audio & webcam feeds across the net - there's an IRC channel #xdevconf on the freenode network for feedback and side commentary - it's displayed on one of the four display walls at the front of the meeting room, so we can all see it. Of course, pretty much everyone here has laptops so we can all login directly to comment, check out websites people refer to during their presentations, take notes, hack on code, check e-mail, or do whatever. I've never before thought I had much use for a laptop, but I'm finding the one I borrowed to be invaluable. (Of course it's running Solaris 10 x86, with Sun's GNOME and Mozilla 1.4 releases, all running on top of the X.Org X11R6.7 server.)

There is very much a Linux orientation, though Solaris has been acknowledged. I have resisted the temptation to yell out "But Sun Ray already does that!" at various points, such as during Jim Gettys wishlist talk when he mentioned how fat clients are so expensive to manage and business really needs thin clients, or that it would be nice to have secure session mobility much like Sun Ray's smartcard hotdesking. Stuart did mention Sun Rays during his talk, and how the usage model is much different than the typical Linux fat client desktop, but shares some similarities with what the Linux Terminal Server Project is doing. I also noted some similarities between the model Keith Packard suggested for a frame buffer driver, especially exporting multiple devices for a multi-head frame buffer, with the frame buffer driver Solaris/SPARC has used for years, but I don't know enough about the model to comment. They are proposing it as a topic for the upcoming Ottawa Linux Kernel Developer's Summit so perhaps we can get them some info about the Solaris SPARC frame buffer driver models by then to see if it would help.

The security talk yesterday was interesting - the SE Linux X work sounds in many ways similar to the Trusted Solaris desktop, but it also appears to go beyond what TS does in some areas. (Unfortunately, I don't know enough about either TS or SE Linux to say much for sure.)

The big splash of the conference was definitely this afternoon when both Sun's Project Looking Glass and the Croquet project presented glimpses into the 3-D desktop future. Both were well received, and seemed to complement each other - LG covered more of the infrastructure to provide a 3-D desktop, while the Croquet presentation skipped the implementation details to show their ideas of navigating a 3-D world much like a game of EverQuest or Doom, but for collaborating with others and doing work. While no one was quite sure how to make 3-D more than eye candy, there seemed to be general agreement that providing the infrastructure was a good thing so that people could start experimenting with how to use it. One thing I saw during the Croquet demo seemed very interesting - a spreadsheet which you could turn at an angle and watch bars pop out of every cell representing the value in that cell, making it an instant 3-D graphical representation of the data.

The two questions asked over and over again of the LG team were "When can we get this?" and "Will it be open source?" Unfortunately, the team couldn't give complete answers to those questions since the details are still being worked out, but they did assure everyone they were trying hard to get good answers soon and to work closely with the open source community.

Posted at 09:26PM Apr 29, 2004 by Alan Coopersmith in X11  | 

Blogging beyond the firewall...

Here goes...my first post to the blog outside the Sun firewall. I've been blogging a little bit inside the Sun firewall but haven't really gotten into it that much - it's hard to think of many things that I think many other people would find interesting (at least that I can talk about publically), but it seems from talking to others there may be a few more than I think. We'll see how this goes. After 5 years of posting to Usenet from a non-Sun address to avoid any issues with people claiming I'm misrepresenting Sun it's nice to see the pendulum swinging the other way, with management realizing that a dialogue with customers and the community can be a good thing and actually encouraging it, despite the risks of engineers saying things that marketing and legal may wish we hadn't. The timing is actually nicely coincidental with my attendance at the X Developer's Conference so at least I'll have some interesting thoughts to get started with before returning to the daily grind next week...

Posted at 07:44PM Apr 29, 2004 by Alan Coopersmith in General  | 

At the X Developer's Conference

[Originally posted to my blog inside the Sun firewall and migrated later to blogs.sun.com]

After spending way too many hours in the airport yesterday, we finally got to the hotel in Cambridge last night at 11:30 pm local time. Our flight was suppossed to leave SFO at 7:20 am - but the plane was an hour and a half late coming in from LAX since they had to change the tires down there. After loading us and taxiing out to the runway, we sat there for almost an hour before they decided there was an unspecified "maintenance" problem that was so bad they had to take the plane out of service and cancel our flight. The passengers were redirected to other flights - Deron got a connection through Dallas/Fort Worth that left SFO around 10:30am PDT and got to Boston around 10:30pm EDT, while Stuart & I got put on the next direct nonstop to Boston which left SFO at 2:15pm and got to Boston only a half hour after Deron's flight. Of course, they didn't have the luggage moved off our old plane in time to catch Deron's flight, so we met him in baggage claim waiting for his luggage to come in on our flight.

Fortunately, the hotel is right next door to the HP Labs facility hosting the conference so we were able to easily find it this morning. I've been able to finally place names with a number of voices from the various conference calls and have had a couple interesting discussions during the ample time set aside for mingling/hacking/etc. The talks this morning were audio cast on the net and via dialin/concall, as well as being transcribed into the #xdevconf channel on the freenode IRC network. http://www.freedesktop.org/Software/XDevConf has more info on the various broadcasts.

Update: Notes are also being published in Federico Mena-Quintero's blog.

Posted at 12:02PM Apr 28, 2004 by Alan Coopersmith in X11  | 

The Fifth Sentence

[Originally posted to my blog inside the Sun firewall and migrated later to blogs.sun.com]

Out on planetsun.org this weekend are posts by Tim Bray and Dave Edmondson, about a "silly internet game", which you play as follows:

1. Grab the nearest book,
2. open it to page 23,
3. find the 5th sentence,
4. post its text along with these instructions,
5. point back to where you got the idea so that we can follow the threads.

Since I was sitting at home by my Ultra when I read this on planetsun, the nearest book was the one on (or rather, in) the hard disk by my knee, and given the topic of that book, I decided to post it both here and to comp.unix.solaris (with the author's permission) - so for your viewing pleasure, the world broadcast premiere of the fifth sentence of page 23 of Solaris Systems Programming by the one and only Rich Teer (who previously worked for Sun in the UK):

The -a flag tells id to show us all the groups we belong to (interestingly, id -a will list up to 32 groups regardless of the setting of ngroups_max).

Coming later this year to a theater, err, bookseller near you! You'll have to pick up your own copy to find out what comes after that gripping and suspenseful buildup. (Though given the size of the review draft I have, you may need a friend to help you pick it up.)

Posted at 10:44PM Apr 24, 2004 by Alan Coopersmith in Solaris  |  Comments[1]

Going to Boston

[Originally posted to my blog inside the Sun firewall and migrated later to blogs.sun.com]

To top off a hectic month, I'll be going to the X Developer's Conference next week in Cambridge, Mass. I'm looking forward to the break, but can't imagine the backlog when I get back. It's important to go though to make sure we have a voice in the roadmap for future X technologies and to make sure the community knows we're serious about being a part of the new X.org open source community. It will also be nice to finally put faces to all the voices from our con-calls and e-mails from putting the X11R6.7 release together, as well as to meet some people whose names I've seen for years.

A couple of people (Deron Johnson & Hideya Kawahara) from our Project Looking Glass team will be there to present our Project Looking Glass plans and start working with the people who have similar interests, such as Keith Packard and his Translucent Windows project which needs much of the same X server infrastructure. Other Sun people on the attendee list include Stuart Kreitman from my group, who was one of the core organizers of the new X.Org Foundation and is an member of the Interim Board of Directors; and Erwann Chénedé of Sun's JDS/GNOME team.

Besides the Project Looking Glass session, the sessions I'm most interested in are the security sessions - including a talk from one of the developers of the NSA SecureLinux work on "Fine Grained Access Control for X" and from Jim McQuillan and Ted T'so on "Network security and X". I'll try to blog here as much as I can from these sessions - especially if I can arrange for a laptop in time and get VPN on it. They are also supposed to be setting up audio feeds for those who want to listen but can't be there in person - I'll post info on those when I get some.

Posted at 05:27PM Apr 21, 2004 by Alan Coopersmith in X11  | 

X11R6.7.0 is released

[Originally posted to my blog inside the Sun firewall and migrated later to blogs.sun.com]

Keith Packard just announced the release of X11R6.7.0, the first release from the new X.Org Foundation (the latest in a long line of groups maintaining X since MIT formed the original X Consortium). This release brings together X11R6.6 with everything in XFree86 4.4 that's not covered by the new XFree86 license (more about that in a bit).

I'm excited by this both because of my personal involvement (I wrote the IPv6 support and was part of the release team, fixing bugs and getting the tree ready for release) and because of the completely different way this release was built compared to the previous releases.

The big change is that it is the first release developed in a truly open manner - where CVS and bug database access are open to the public, instead of restricted to the paid members of the industry consortium, and indeed a growing number of community members and volunteers are participating from outside the tradtional Unix vendor companies who previously made up X.Org.

It also comes on the heels of great dissent in the XFree86 community, starting with the ejection of Keith Packard (author of the Render extension, Xft, fontconfig and other popular X projects) from the XFree86 Core Team and spreading even wider with the recent adoption of the new XFree86 1.1 License which many Linux & BSD distributions object to (some claim it's incompatible with the GPL, others that the requirement for giving credit anywhere you credit anyone else is to onerous to comply with). Because of this, many of the distros are planning to switch to the X.Org software for their future releases - RedHat has already switched Fedora, and SuSE is looking at switching.

Posted at 12:03AM Apr 07, 2004 by Alan Coopersmith in X11  |