As of today, I have resigned from my current job at Novell. I worked for Novell as a consultant from 2004-10-04. There are two main reasons for my decision.
Disagreements with Novell's leadership
I don't share some of the technological decisions and positions that Novell's leadership has; I would be doing some things differently.
I did try to voice my opinion internally and make a difference, —at some point I even became one of the two co-leaders of its Open Source Community of Practice in order to do this—, but I found it very difficult for the thoughts of regular employees to reach Novell's top executives. I suppose this isn't something particular of Novell but happens in most organizations its size.
These disagreements, along with the, not suprising, lack of a channel to participate in the decision-taking process, were very frustrating for me.
The following are some notorious examples of things I would have done differently. The list is not exhaustive.
The only reason I'm writing this list is because I think some people within Novell may find my opinions useful. This should not be taken as criticism of Novell's current policies: there is some incredibly intelligent people currently working at Novell and I may be missing some of the information they have. I hope the list turns out to be useful for Novell.
OpenSUSE's «openness» (or lack thereof)
I would have handled OpenSUSE very differently. I think it was mostly just a publicity stunt, where a lot more «openness» was implied than was actually there. I would have pushed for a far more «open» model, one that would, among other things, allow third parties to maintain official packages.
See this thread for some details.
Update: Andreas Jaeger kindly pointed out in the comments that they are now working precisely in this line. As you can see in the thread I quoted, this was definitely not the case back in 2005: over 10 people replied to my message and not a single one agreed that it would be a good idea (someone even claimed that dropping the idea of a centralized distribution on which everyone can get changes / packages was “fundamental for the future success of openSUSE”). It is great news to see that OpenSUSE is now moving in this direction.
Patent cooperation with Microsoft
I wouldn't have done the “patent cooperation” part of the agreement with Microsoft. It obviously was a workaround to the spirit of the GPL license.
As a person who has decided to use it to license programs to which I have dedicated a great amount of time, I did not like this. I chose that license because its spirit matches very closely my intentions with regards to who may use my work and how.
While it may seem the most important, this may be the weaker of the three disagreements I'm listing. I can see Microsoft pushing for the patent cooperating agreement and refusing to do the entire agreement without it. The entire agreement is indeed bringing a lot of new business oportunities to Novell allowing its distributions to enter many important organizations (for example, it has been announced that Wal-Mart will be using Novell's server distribution and, when this happens, it will be huge). So my guess is that Novell executives had to compromise on that part. It is easy to criticize the bad aspects of compromises but the other alternative, in this case, seems to be no agreement whatsoever, good-bye opportunities, not a very good option for Novell.
Focus on services
Novell sees itself as a traditional software vendor, with GNU/Linux being just another product, its consulting arm and services in general just minor focus areas. It is trying to sell GNU/Linux as if it was just a regular software product, such as GroupWise or eDirectory, using its “subscription” model. It wants to go as little as possible beyond creating the software and providing it to partners: let those partners be the ones to sell it and provide services to the final customers. The less direct contact Novell has with the customers, the less «touch», the better.
I think Novell's business model around GNU/Linux is not going to work: I doubt it will be able to earn big money selling its GNU/Linux distributions as its traditional, proprietary, products.
To charge a significant amount, Novell would need a big differentiator from other distributors. However, if the differentiator is based on free software other distributors will simply copy it and the differentiating characteristic will cease to be such. Ergo, Novell would need to add significant non-free components. However, following this route and making its distributions depend significantly on non-free software would also mean giving up one of its main differentiators from Microsoft Windows as well as giving other GNU/Linux distributions an important differentiator from Novell's. Novell just can't afford this.
Of course, there is also the possibility of creating differentiators based on things other than software, such as the availability of commercial support or a large number of certifications for devices or third-party software. I can tell you Novell has, consciously or not, being trying to create many such differentiators. While I think these are the main differentiators Novell has from other distributors, I don't think they work very well, not well enough to justify a large price, specially when your focus is to sell GNU/Linux as just another product and sell very little services around it directly. Even if Novell manages to create a large ecosystem of strong partners, as it starts to increase the price of its distribution, it will start to create incentives for these partners to also work with other distributions. Once you've learned one distribution inside out, you can very easily learn another one. Furthermore, I think the case of Ubuntu shows that widely available commercial support is not that critical a factor for the success of a distribution, even in the corporate sector. Believing otherwise is just wishful thinking.
So I think the only way Novell would make big money on GNU/Linux would be to focus on services around it. Sadly, I perceive it to be going in the entirely opposite direction. I suppose one of the reasons for this is that the margins for selling services are significantly lower than the margins Novell usually enjoys when selling traditional proprietary products.
Novell could simply accept that it will not be making big money from its GNU/Linux distributions directly and will simply use them as a tactic to get their feet in many corporations to which it can then sell its non-free products. I don't think this would work very well because many of the corporations migrating to GNU/Linux are doing it because of all the practical advantages of switching to free software; these corporations will, in general, not be very prone to buying non-free software around GNU/Linux. And even if it actually worked, I think Novell can do better than this and actually make money directly from selling GNU/Linux if it were to strengthen its services around it.
Tired of consulting and traveling
The second reason why I resigned was far more personal: I grew tired of working as a consultant and being constantly traveling.
Software Engineer vs Consultant
I did work in some very interesting projects from which I did learn a lot. I realized that, while in the past I tended to focus mostly in technical skills, soft skills are extremely important and are not as easy to master as I thought. I worked with some very intelligent people from which I learned a lot.
But I really enjoy writing software so I decided to move to a different type of job, one as a software engineer. My job as a consultant involved some things that I, personally, did not enjoy and that I won't have to do in my new job. Also, it seems that my new job will have many challenges that will be far more interesting than the most interesting challenges I've had as a consultant.
Traveling
As for traveling, I got to know a lot of places. I visited Venezuela, Perú, Ecuador and around ten states in the United States in around two years. I lived for periods of over 5 months in Arkansas, Florida and Venezuela. I did have a lot of fun traveling.
However, medium-term trips come with some headaches and, one after another, they wear you down. You begin to feel disconnected, the idea of «home» begins to disappear: it certainly isn't the one city you just got to, where you don't know anyone; nor is it the city you've been living in for 3 or 4 months where you now have a few friends but which you'll be abandoning in a few weeks, probably never to return; and the city you'll be returning to, —which you have visited only for three or five weekends in the last four months—, also gradually loses its «homeness», as all your friends carry on their lives without you and you have to make a conscious effort to keep in touch and keep the spaces you shared with them alive.
In the new job I'll be taking I may have to travel for brief amounts of time but I will have a permanent home where I will be most of the time.
Current status
So that's my current status.
I'm very grateful for the opportunity I had to work in Novell. As I said, I did learn a lot of things and I met some very intelligent and nice people. Novell was a very good employeer, overall. However, for the reasons I explained, this was time for me to move on.
I was very lucky: just when I was starting to tell my close friends that I was thinking of looking for a different job (shortly before I posted the look-for-positions piece with advice for colombians), I was contacted with a very interesting offer. I had to take some interviews and I was accepted. I will now be working as a software engineer in Switzerland. One thing that makes this offer very interesting is that I will likely end up working with some very good friends of mine.
Right now I'm taking some quiet time in Colombia, doing all the paperwork required for my work visa and my relocation.
So long, Novell, and thanks for all the fun.
Discussion (20)Last update: 2007-11-07 (Rev 13075)