More about Directory Server and OpenDS Margin Notes

Pierre and Gilles have mostly rewired our central servers and labs to the new switch in Sun's Grenoble engineering center.

Here is the before picture:

And here is what it looks like now:

Why should anyone care? Well, now we have a 10Gbit backbone between labs, with 800Gbit throughput in the central switch. Pierre ran tests showing that a machine with 1Gbit Ethernet now actually can get 1Gbit throughput to a machine in another lab and network. That means we no longer have to have all equipment for a big test physically located in the same place.

That freedom is a good thing. My expectation is that we will soon be trunking 4 x 1Gbit Ethernet for some of our heavy load tests with OpenDS, for example. 

Luke Donnelly sent this around a little while ago. Interesting for somebody living and working here in Grenoble, France. In a blog entry from last fall about Yahoo! doing engineering in Grenoble, you can read this:

"We operate in a very competitive, multinational world," said Linwood. "We do development all over the world, but it's important to find locations where we can find talent, work with the government and grow our business in an environment that is friendly to that. These changes have made it much more attractive as a company [to invest in France]."

This year the incentive for tech companies to do R&D in France has increased. The French are giving companies credit for a big part of their R&D costs, whether or not those companies have made enough to have to pay tax on revenues in France. Maybe this sort of busts the myth of old Europe having higher costs in the end.

BTW, despite all the show this year, the 35-hour week is still a myth, too. (Sometimes I wish it weren't. ;-)

Today is Software Freedom Day 2008. The celebration's seem to be going on around the world.

Over here near Grenoble, France, we will be celebrating with the latest build of OpenDS of course. But I am also planning to teach my son to apply at least the security patches on his Ubuntu desktop without Dad's help ;-)

This was a bit trickier than Ubuntu in VirtualBox on OpenSolaris. The screen resolution was resolved by the Developer Preview 2 install inside VirtualBox, but the OpenSolaris VM had no networking. How do you download missing stuff when you have no network access?

You get the .iso for the ae network driver from Alan Burlinson's blog. I do not have bunzip2 on Windows XP, though you can probably get binaries somewhere. Instead I unzipped the .iso on another system, stopped the OpenSolaris VM, and then mounted the .iso.

Next I followed the directions in the Readme file for the driver ( more or less ;-). After a reboot of the VM, the network seems to work fine.

Giving a beta a try is easier with VirtualBox. Beta within beta, in fact, since VirtualBox seems to be in beta on OpenSolaris.

All I had to do to get started with the install was mount the beta Ubuntu .iso as a CD-ROM. What's not working yet? Looks like screen resolution is stuck at 800x600, and USB does not get cabled through correctly.

Last time I tried to do the same thing with VMWare, I had to find someone with VMWare workstation. Kind of expensive if all you want to do is give some beta software a whirl.

                  OpenSolaris Developer Preview 2 snv_79b X86

                         _        _     _         _ _                
           _ __ _ _ ___ (_)___ __| |_  (_)_ _  __| (_)__ _ _ _  __ _ 
          | '_ \ '_/ _ \| / -_) _|  _| | | ' \/ _` | / _` | ' \/ _` |
          | .__/_| \___// \___\__|\__| |_|_||_\__,_|_\__,_|_||_\__,_|
          |_|         |__/                                           

(Published from my Tecra M5)

Lana sent along links to a couple of articles on C|Net, Open-source strategy: Documentation = dollars and Why people write documentation.

The former article explains that for Ian Howells, Alfresco's Chief Marketing Officer, folks both reading their documentation and evaluating their software are great prospects. Also:

People don't visit a software company's website to read about the executives. They visit the website to get information on the software.

The latter article claims community building is the top reason to write documentation. Community building wins over personal growth, gratitude, reputation, and other worthy contenders.

One of the reasons I came to Sun was to do structured documentation. Jon Bosak was working at Sun when he "fathered" XML. How could one get any closer to the center?

When I finally got a chance to go out for something to eat with Jon Bosak, he seemed more interested in General Robert's Rules of Order and electronic democracy.

Structured documentation is fantastic when you want to process and publish the content through different systems in different ways. With Directory Server Enterprise Edition, we run a cron job nightly that builds our docs checked out from the source tree, looks for issues in the production process, checks that all the links point somewhere valid, and posts the very latest on our internal site every morning. We can generate PDF, HTML, JavaHelp, now even WikiText. We are way ahead of what we used to do years ago, painstaking generating everything by hand.

There are at least two problems with this approach, however. One problem lies with a characteristic of structured documentation itself. It takes some expertise to write structured documentation. Not much expertise, but enough to prevent people from getting involved casually. Another problem comes from the speed of early release. Even in the middle of the night, tomorrow morning is a lot later than right now.

With the OpenDS Wiki, we had a completely new set of docs to write. We decided therefore to take the plunge, sacrificing the use of our great structured doc tools to the larger goals of lowering the bar to authors, and of dropping publication lag from overnight to instantaneous. Even better, since OpenDS development is out in the open, the Wiki is constantly crawled by your favorite search engine (as are the mailing lists). You no longer have to be in the know to get access to the latest content.

To what extent do the benefits of traditional documentation outweigh the benefits of having documentation on a Wiki? Feel free to answer in the comments for this entry.

One of the biggest benefits with the OpenDS Wiki has been the high volume and high value of community contributions. Wiki technology makes it simple, almost tempting, for folks who know their subject matter to add good content. Perhaps because of the public nature of all edits, the signal to noise ratio stays high, too.

Even so, we want to put the quality bar for official documentation as high as for the code itself. That is why we apply a two-phase documentation review process, similar to a code review and testing process.

Will that ever be good enough? Maybe not.

When your operations depend on the reliability and availability of your directory service, you want one throat to choke, right? Any community has more than one throat to choke.

Our aim today is to continue with wikis as the community source of the latest and greatest documentation. We then expect to bring refined and validated content from that format into official product documentation. Finally, we plan to package that up in such a way that you know exactly whose throat you get to choke when you buy support.

If only the middle step were easier to automate. It would be nice to have an open source WikiText-to-OpenDocument-format conversion tool.

Calum Benson suggested using Yahoo! Pipes to aggregate the writer blogs.

I have a Yahoo! Pipe feed that gives you the last 50 entries with titles preceded by author identifiers so you can tell who wrote what. Seems like that should have been simpler than it was. Or maybe I am dumber than I realize.

Update: JohnD has gotten an aggregation created for information products blogs at http://planets.sun.com/SoftwareInfo/.

A picture is worth a thousand words. A video consists of thousands of pictures. Thus one video equals at least a million words. Now that you have thought that through, check this out.

Hours of footage of the sys admin typing configuration commands wound up on the cutting room floor.

JohnD noticed a bug in this blog. By default, I was supposed to moderate comments before they could appear. Also by default, I was not sent notification about any comments.

How was I supposed to know? RTFM ;-) The answer was there, halfway down a web page full of settings. All I had to do was read all the fine print very carefully and thoughtfully before doing anything.

Update: Took some time today finally to respond to questions in your comments. Thanks for those.

JohnD started a list of all the writers at Sun who are blogging. I just added the list of writers to my bookmarks.

A long time ago, somebody coming from Netscape wrote that Sun revolves around email. We mention that email can be archived and then searchable, which is true. Without email archives and search engines, lots of software would be undebuggable for the mere mortal.

Yet deep down we most like email because it's asynchronous. We like blogs for that, too. We like quiet uninterrupted problem solving. We punish people who like to be interrupted by promoting them to management.

But perhaps we need to build in a little more interrupt handling time. Andrew C. Oliver was telling us about using OpenDS. Here in email Andrew's responding to Neil answering on behalf of the rest of us.

> - The IRC channel is completely dead because no one knows anything about 
> it.  I didn't know it existed at all until Pat Patterson asked me about 
> it yesterday (possibly as a result of a question that you might have 
> asked).  I have no idea who set it up, but it's not really affiliated 
> with the OpenDS project.  We would prefer mailing lists for this kind of 
> communication, since they are archived and searchable, and since you're 
> more likely to get a larger audience and therefore a more timely and 
> potentially more accurate response.
> 

Just my 2c.  I find real time communication is more useful for some 
things.  However mainly it is a useful community building tool.  At 
present you're doing this based on geo-location and internal company 
communications.  It is much easier to archive IRC than 
intra-cubicle-chat but has a better bandwidth/comprehension factor than 
email for some things.

I've also basically learned that every open source project will have a 
channel on freenode whether they like it or not :-)

You can read what Andrew has to say about OpenDS at his blog on buni.org.