Now the Desktop Linux Summit update. The show is a little shorter than some (only 2 days) and has three tracks, but the tracks were pretty jam packed. No longer at the fair grounds, DLS has upgraded to the Grand Hyatt. Doc Searls had an extremely interesting and funny presentation on the first day and I was lucky enough to be able to catch up with him and a bunch of other people (Arne, Geo, Steve, etc) for a couple of drinks later. You often learn as much after the conferences as you do during 
Also on the first day was a preview, by Nat Friedman, of what Novell has in store for the desktop. Some nice eye candy on the way, that's for sure. The second day held the main Keynote by Geoffrey Moore, where he covered Linux on the Desktop in relation to the chasm. I didn't agree with 100% of what he said, but he is an outstanding speaker and raised some excellent points (ie. it got you thinking). Ian Murdock and Michael Robertson also spoke on day two. Probably one of the most anticipated talks of the summit came when Rob Enderle got on stage (Google if you're not familiar with Rob). While he's still certainly a bit misguided on some things, I have to admit that he actually made some points on multiple different issues (I'm as absolutely shocked as you). Either he now has one toe in reality, or he was afraid of getting lynched (which Arnie seemed ready to do). All in all a great show and one I'm proud that have the LQ name associated with. I hope to attend again next year.
–jeremy
DesktopLinuxSummit, Linux, Open Source, Enderle
Archive for April, 2006
I should probably post a Desktop Linux Summit update first, since that event happened first, but I'm still at the MySQL UC and have a couple minutes before the next session. The conference has been very good so far. It opened with Martin Mikos, the CEO, having some very good things to say about the future of MySQL. They have an alternative storage engine on the way for those worried about the InnoDB future at Oracle and have just renewed their InnoDB contract with a multi-year agreement. The two of these in combination may quell some peoples fears. Next came the Tim O'Reilly radar speech, which is great no matter what conference you see it at. Some presentation issues aside, this one was no exception (and I have to say, Tim did a very good job of covering things for about 15-20 minutes with no slides at all).
The sessions have all been interesting and I learned a few things I should be able to apply to LQ (I've been attending mostly performance and scaling sessions). While asking one speaker about his experience with FULLTEXT, Monty actually stepped into the conversation and took about 15 minutes to literally dig through the MySQL source and see when some changes were applied. The bad news in that the change he was after is present in the version we are running. The good news is that the recently announced falcon storage engine should have a FULLTEXT engine that rocks. After explaining the situation a bit more, we decided that in the mean time a 64-bit MySQL with a ton of RAM is probably the only viable solution. At how many companies would the founder of the company sit down with you (at a conference no less) and start digging through code. Absolutely awesome – thanks Monty! The rest of the session lineup for the day looks interesting and I should have time for at least one BOF before I have to catch my flight home. DLS update coming soon.
–jeremy
MySQL, OReilly, Open Source, Linux
Kind of a last minute decision, but I'll be attending both the Desktop Linux Summit and the MySQL User Conference next week. LinuxQuestions.org is a sponsor of both. Getting to both these on such short notice was both easier and cheaper than I'd have thought, but it does mean I am doing a whole lot of flying. Should be well worth it though as both look extremely interesting. Unfortunately, I wasn't able to book either conference hotel, the Hyatt was booked solid. If you'll be attending either event and would like to connect, contact me ASAP. I'll be at DLS on Monday and Tuesday, then MySQLUC on Wednesday. See you in San Diego/San Jose.
–jeremy
Desktop Linux Summit, MySQLUC, MySQL, Linux, Open Source,
I've covered this topic a bit before and it certainly fits into my recent discussion of what are we as a community willing to give up if we want to see mass adoption (latest part of that discussion here). A recent article, entitled New Linux look fuels old debate brings up a long standing point of debate. Should Linux allow and use proprietary drivers? This debate covers a wide spectrum of topics and many of the participants are quite ardent in their view. On one hand, one of the founding principles of Linux is Open Source and freedom. Why should closed drivers be allowed then. They taint the kernel (from a GPL perspective at minimum), allow very low level access to your machine without you being able to see what's going on and make diagnosing problems extremely difficult. Things like a Sony rootkit for Linux would now be possible also. Besides that, they just seem to be fundamentally contradictory to the aforementioned founding principles. On the other hand, if this whole thing is really about choice, shouldn't vendors have the choice to distribute closed drivers and consumers have the choice to use them? Won't free markets decide what's best? Seems like both sides have a point, doesn't it.
The reality, as usual, is that it's just not that simple. There are many factors at play here and to explain them all would be much to time consuming. The first thing that has to be settled is whether or not closed source kernel drivers are even legal. I've seen yes as an answer and I've seen no as an answer, both with logical arguments. Looking at the wording of the GPL, I'd guess probably not legal, but if Linus allows them (and in fact the entire kernel community has indeed allowed them thus far), I'd find it hard to believe that a court would make a ruling against them now. As muddy as the legal issues are, I'm going to sidestep the whole topic and leave that to the lawyers. To me, this is more of a “can Linux allow proprietary drivers” and not “should Linux allow proprietary drivers” question anyway.
So, that brings us back to “should”. Some predict that doing so could equate to a Linux doomsday, while others say not allowing them will result in Linux being a fringe OS forever. So, who's right? As usual, I don't think there is a right or wrong here. It depends on what your vision for Linux is. For most of us that have been in the Linux community for a while, I think the answer is somewhere between binary drivers are tentatively bad but borderline acceptable to binary drivers are insidious. If you're a marketing type or someone who things Linux should do anything it can to get 80% market penetration, then you probably answered that binary drivers should be welcomed. Much different visions. But the kernel is too fundamental to Linux to have it both ways. One side has to give. It's another question of will what has gotten us this far be abandoned (or partially abandoned) in order to get us over the next chasm. Can we convince a company like nVidia, who's current potential revenue for Linux sales pale in comparison to potential closed OS revenue, that Open Sourcing their driver really is better for them? Or, will consumer demand for a working out of the box nVidia experience force one of the major distributions to cave and start shipping? Will shall see.
–jeremy
Linux, kernel, Open Source, binary drivers
While it's been suspected for some time now, Microsoft has publicly admitted that they silently fix some vulnerabilities with absolutely no disclosure at all. From the article:
“We want to make sure we don't give attackers any [additional] information that could be used against our customers. There is a balance between providing information to assess risk and giving out information that aids attackers,” Reavey said.
We all know that security through obscurity really doesn't offer any level of protection at all. Don't get me wrong, I'm all for responsible disclosure, but no disclosure at all is just not acceptable. The wrong people will expend the effort to reverse engineer the patches and figure things out. The people that suffer are the helpless Windows administrators. Microsoft has created an admin culture where only the patches that impact an environment are applied by most Windows admins. Part of this is a result of so many patches gone bad. But, if the security bulletin for a patch says it fixes one thing, but really also fixes 4 other things silently, you never know what you're vulnerable to. Now, I'd recommend installing all patches of course, but that's just not reality for most of the Windows world.
This brings up another topic though. A while back I posted about the year-end vulnerability summary that showed Linux/Unix had more vulnerabilities than Windows. So not only did the report include multiple counts for single apps and apps that are not even included in base distros on the Linux side, it also didn't count vulnerabilities that either Microsoft never fixed or ones that they silently fixed during other patches. The worst part of this is that Microsoft uses reports like this in their marketing. So, they don't ever fix some vulnerability, silently fix other and then claim that hey have less vulnerabilities. All in the name of “customer best interest”. Yikes.
–jeremy
Microsoft, MSFT, disclosure, security, Linux, Open Source
So it appears that Windows Vista is going to do a piracy check and then disable Aero if you don't have a licensed copy. That's right – it won't refuse to run or even cripple functionality, you just won't get translucent windows or animated flips. You see, Microsoft won't admit it, but they need piracy. It's one of the ways they keep the market share that they have. Keeping that market share is extremely important to their business plan. Of course, they don't want piracy to be completely rampant, so they do things like this. Considering most pirated copies of Windows are the corporate versions that don't phone home anyway though, I think there is more to it. By disabling some features like this, it makes the OS seem less like a commodity and more like something special. Once the OS level has been commoditized (something Linux is driving toward) there is just no way Microsoft could charge what they do for Windows. With the trend toward commoditization consistently growing and the Windows market share slowly being chipped away at, you realize how important Vista really is. Add in the fact that Mac hardware prices are falling along with the improvements in virtualization and emulation and things are looking worse and worse for Microsoft on the Windows front. In the end, the average consumer should benefit from this greatly. Not only will they have real choice for the first time in a long time but all OS's will need to remain more competitive. Most Microsoft products have stagnated recently since they didn't need to improve to survive. Luckily, those days are swiftly coming to an end.
–jeremy
Vista, Microsoft, MSFT, Windows, Aero, Open Source, Linux
Taking a quick look at my last post (which is tagged LinuxWorld, I realized that I never posted a LinuxWorld wrapup. LQ once again exhibited at LWE in the .org Pavilion. We've been participants in numerous cities from New York to San Francisco to Boston to London. Exhibiting at LWE is always a fantastic time. The .org Pavilion always has a very fun set of both visitors and exhibitors. This year was no different. In fact, the “Best of Show” winner OpenSuse with located directly next to us. The LQ booth was a little less staffed than normal for us. This was the result of a last minute issue that came up for a traveling mod. The show was a little quieter than normal though, so Robin and I had no problems manning the both with just two people. Asking around a little, most (if not all) exhibitors said the show was quieter this year than last, and almost all attributed it to the location (which was BCEC this year and Hynes last year). The new BCEC is very nice, but the location of Hynes is much better. Back to the show itself, it started off with a bang when the Unisys booth literally caught on fire. Things settled down from there though and the rest of the expo went as expected. One interesting thing to see after exhibiting at these for a fair number of years now is the shear number of people who are either LQ members or at least remember getting help from the site. When we first exhibited in New York, the most popular question, BY FAR, was “what is LinuxQuestions.org”. That is no longer the case by a long shot. The shear number of people who come over to the booth just to shake our hand and say hello is staggering. It makes it all worth while (and then some). As usual the after expo festivities were in full swing in Boston. We didn't make the SugarCRM party unfortunately, but some of the Fedora guys said it was a great time. I got to chat with the TLTS guys quite a bit and they're a great group. All in all another great LWE. I hope LQ is once again invited to San Francisco, if we are I'll see you there. On the conference note, I have confirmed my trip to San Diego for DLS, so if you'd like to connect there, drop me a line.
–jeremy
LinuxWorld, LWE
It looks like Jeff Ayars from Real had the following to say at LinuxWorld in Boston:
“The consequences of Linux not supporting DRM would be that fixed-purpose consumer electronics and Windows PCs would be the sole entertainment platforms available,” Ayers said. “Linux would be further relegated to use in servers and business computers, since it would not be providing the multimedia technologies demanded by consumers.”
A couple comments here. First, Linux is not nearly currently anything that would even come close to resembling a consumer entertainment platform. 9 out of 10 distros still don't support MP3 out of the box. Before you chime in with shouts of livna (which is fantastic and I use and recommend), note that we are talking consumer here. If you think the answer to playing the most widely supported audio format is “first, enable this yum repo and then install the following packages”, then you're clearly not the average consumer. So that brings us back to a topic I have talked about on this blog time and time again. What are we going to give up in the Linux community to get that much talked about “mass adoption”. My guess is, you're going to find out very soon. I'd say 12-18 months at the most and we'll start to see some of the myriad decisions start to unravel. I've said it in the past and I'll say it again. I'd prefer not to hit 90% market share if it costs us much of why most people think Linux is good. What market share do I think is the max if we don't conceded on some of these issues and what market share is acceptable for the major Linux vendors out there? The former is a question I've been thinking a lot about lately and one that I had a lengthy discussion about at LinuxWorld. I'm still thinking though, so I don't have an answer yet. The latter I really have no idea about, you'd have to ask them.
So, that brings us to the next topic. DRM. Like most others, I really don't like DRM. I'm pretty sure consumers are not clammering for DRM. What they want is the content, and the media companies currently seem to be insisting on putting DRM on that. One problem with DRM though, is that it barely even slows down piracy, but is a major hindrance to people trying to do legitimate things in a way that the content industry may not have thought of. Something as simply as purchasing a song in iTunes and playing it in Linux for instance. The reality though, is that not only do most people not care about DRM – they don't even know what it is! I think that will soon change though. Why? Let's break people down into four basic groups. This may be a bit of an oversimplification, but I think it will illustrate my point.
1) The average user
2) The person who will pirate everything, for no other reason than to do it
3) The technically savvy crowd who dislikes DRM not because they want to pirate, but because of the privacy and convenience issues
4) The few people technically skilled enough to actually figure out and circumvent DRM and then distribute tools for groups 2+3 to use.
Right now, DRM really only impacts group #3 (which I would put myself in – I have *no* problem purchasing 100% of the software and content I use, but if it doesn't work on my platform of choice I have no problem simply finding something else to use/watch). Group #2 uses the tools that group #4 created and has 300 movies download of which they have probably watched 4. This group should be considered non sequitur for the content industry, since nearly NONE of them would likely purchase any content no matter what. That brings us to group #1, which represents the vast majority of the content purchasing public. Right now, DRM probably has not inconvenienced them too much. However, if the content industry gets their way with some of the proposed asinine implementations, they will be severely impacted. Who will they turn to when their file doesn't play, their movies doesn't load or whatever other content they have doesn't work? If you're in group #3, you're probably smiling, because you know it's you. The content industries are getting greedy now. They are turning record profits (pun intended) but they want more. But their greed will be their undoing. Once group #1 becomes sufficiently inconvenienced that they start listening to what group #3 has to say about DRM, there can be only one outcome. The same outcome that always eventually happens when a company or group tries to do something that is inherently not in the best interest of the masses. Revolution. The sad part here is that DRM could legitimately be used in a whole host of interesting ways. It seems greed my just not allow that though.
–jeremy
DRM, greed, LinuxWorld, media, Real, Open Source, Linux
Sometimes after I blog about something, I click the technorati tags for the post and check out what others are saying about the topic I've posted about. In fact, I did so for the previous post, entitled Red Hat Signs Definitive Agreement to Acquire JBoss. Indeed, I found something interesting. There were more than a couple links to this post, from Jboss CEO Marc Fleury. The post is blank, you might say after clicking on it. Indeed it is, but enter the Google cache. The Sun vs Red Hat, who is more Open Source post contains the following (amongst much more):
* So to me both SUN and RH are open source “wannabees”, or as one of my developers put “open source girly men”.
* Today RH *IS* a proprietary vendor.
* RH is a packager, it doesn't create JACK, it doesn't create Linux, it wraps it up in proprietary shit. And no the contributions that they make don't really count. Linus Torvalds creates Linux.
* RH likes to pretend they are the open source community, waiving its flag. While this used to impress Wall Street it doesn't impress us.
Well, you get the idea. Now to be fair, that was posted a long time ago – “28 Sep 2004″ in fact. While I don't find it odd or disturbing that Marc said any of those things, nor is there any indication that he still believes those things (at least in that exact form) two years later, I do wonder when and why the post was actually removed. I checked archive.org, but they have zero copies of the page. The Google cache says “as retrieved on Apr 7, 2006 03:44:47 GMT”, but I have no idea how accurate that really is. One of the great things in the Open Source community is the transparency. In the days of blogging and the Google cache, it's hard to hide anything anyway. I'm sure if you search on this blog you'll find something I said in 2004 that I'd vehemently disagree with now. We all live and learn after all. But that's a good thing, not something to hide. I still very much believe that the acquisition is a good thing for Red Hat, Jboss and the Open Source community in general…but curiosity does have me wondering about the exact delete date and what the removal reason was. Maybe he was just getting harassed for continually calling Michael “Tielman” instead of “Tiemann”
–jeremy
Red Hat, Jboss, RHAT, Fleury , Open Source
The rumors were wrong. Oracle isn't going to acquire Jboss, Red Hat is. A very interesting move I think, one that is a little more aggressive than Red Hat has been lately. RHAT went up almost 10% on the news. From the press release:
Red Hat will acquire JBoss for approximately $350 million in initial consideration, plus approximately $70 million subject to the achievement of certain future performance metrics. The transaction consideration is composed of approximately 40% in cash and 60% in Red Hat common stock.
I think this is extremely good news for Jboss and its community. Red Hat gets Open Source. They have proven it time and time again. The end goals, not to mention community and culture, of Jboss are also much more aligned with Red Hat than with Oracle. While Oracle hasn't done anything crazy with InnoDB or Sleepy yet, you still have to admit their long term independence is far from secure. With this purchase, Red Hat is taking another step further away from being a software company and closer to being a services one. That's another nice thing about merging the Jboss product offerings into the Red Hat suite – they're already a subscription based model.
Taking a step back for a moment, this move has all kinds of ancillary repercussions. Red Hat has traditionally had IBM as a very strong partner. They now have a directly competing product though, so we'll have to see how IBM reacts. RHEL has always been one of Oracle's preferred platforms. With Red Hat + Jboss will we now see a minor shift toward SLES from Oracle? BEA is already jumping on this one, claiming that they are now the only independent vendor of application servers, “not tied to a particular OS or database”. They've even gone as far as to say they would potentially leverage Jboss and contribute to the project if the IP model changed. In the end, I think this was a great move for Red Hat though, assuming they can limit the Oracle and IBM fallout. This will get their foot in the door at a whole range of high end installations that they just wouldn't have seen before. Now, they just need to convert that foot in the door to more subscription sales.
–jeremy
Red Hat, Jboss, RHAT, Oracle, Open Source, BEA
Subscribe