Raw Matter

(Let's test this planet thingy... Hello everybody!)

Setting up a SLES 10 server at $WORK, for a change. Compared with my usual distribution, this means: LDAP authentication is easily set up, I don't even have to know about all the packages (PAM, NSS, LDAP server, ...), and when I then create a user in yast, it does create a user in the LDAP database automatically. Debian could learn a thing or two about integration here. But it also means a hideous apache configuration schema, with lots of configuration files, some parts of the configuration in /etc/sysconfig instead of /etc/apache2, and config file snippets generated by the init.d script (coming from SuSE 4.x days, I felt right at home. At least yast hasn't overwritten anything I didn't want it to so far.) And it also means that I had to compile subversion from source because I couldn't immediately find SLES 10 binary packages. Of course, I have to admit that Subversion is a truly exotic peace of software that no sane systems administrator would ever use on his server, so I can understand that.

Conclusion so far: SLES10 is a nice piece of software, but some aspects can be called rough edges although I guess many would argue it's just my bias.

Hoi Martin,

Brack in Mägenwil: prices are ok, and I had several warranty cases with them and never had any problems. Call them, return the defective item to them, get a replacement. For dead on delivery PSUs I even had them send me the replacement immediately when I promised I'd send them the defective unit back ASAP. I always ordered online and had stuff delivered, and while you can go and pick stuff up directly, I'd not necessarily expect advice from their staff. OTOH they won't try to sell me something I don't want.

Diversimex is my walk-in-and-chat store when I spontaneously need some parts. Nice chap, sometimes coffee included with your purchase. It's a small store, so they can't have everything in stock, but he knows more or less what's going on, and if you know what you need he'll order anything for you.

(My nanoblogger doesn't do traceback pings, so: yes, this is a replay to madduck's article from Friday.

I wonder why some people just can't live with the fact that being very vocal on mailing lists doesn't have to mean they're part of a majority.

To all those who are “just” continuing their work to help Debian release: thank you.

Ben Chaff argues that Free Software is better than average in terms of security when compared to the proprietary software market, but falls short of the standards that apply for software used in crash-and-people-will-die type of mission critical applications.

I fear I have to cry foul here: developing software for the kind of systems he describes (nuclear power plant controlling, medical equipment, ...) means developing software with a large budget and a limited, relatively well specified functionality. Compared to this, Free Software is mostly written on a tight budget (even though no monetary compensation is awarded to many Free Software authors, it is only fair to call the willingness to invest time into a project a budget here) and often with a heavy focus on extending functionality.

In other workds: Ben is comparing apples to oranges here. Here is the interesting part of Ben's article:

The irony is that our research shows that on average, open-source software is of higher quality than proprietary software. Indeed, open-source projects tend to clump together in the higher-quality range. Proprietary software applications scatter across the quality continuum, but the best ones tend to be considerably better than open source, and customers don't choose software based on industry averages.

While Ben sees this as somewhat ironical, I interpret this fact in a vastly different way: Free Software projects are judged on their merits and not by how much can be spent on marketing. If they don't meet at least some minimal quality standards, they tend to die in their infancy. So, leaving the world where money is available to actually ensure top quality and entering the everyday world of business, where software is deployed after a rather short evaluation period, because a solution is required now and not in two years time, picking Free Software by default could be a safe bet.

So, let's compare Free Software against proprietary software. But not a proprietary software to control a nuclear power plant against a Free Software groupware suite. Let's make the comparisons that actually come up when people deploy software: MS Internet Explorer against Mozilla Firefox. PostgreSQL and MySQL against Oracle and MS SQL. GCC and Eclipse against whatever proprietary compiler and development environment you get when you buy an ARM-based experimental board. Diebold voting machines against the voting software used by the Debian project.

And, separately, answer why there are no Open Source jet engine controller tools or Open Source nuclear power plant control software packages. For me, the answer is relatively obvious: Free Software works when a community of people with related problems can work together to implement a solution. Yes, some Free Software projects are started by a single person or company, but no Free Software project has survived without a surrounding community. And you just won't deliver a nuclear power plant to me just because I'm interested in the technology...

As an aside: I had to laugh out at this bit:

Because of the higher average quality of open-source software revealed by our research, we strongly believe it can cross the chasm into mainstream use.

So Google running entirely on Linux is not mainstream? Microsoft using a TCP/IP software stack with Open Source origins is not mainstream? Firefox grabbing 10 per cent of the Internet Browser market is not mainstream? Apache holding 60% of the web server market is not main stream?

Update (2006-11-06): Fuzzy feelings time: this article, together with the one I'm responding to, hit Slashdot, was featured on Linux Today and even translated into Italian by the folks at Doxaliber.