Raw Matter

Matthew Newton nicely wraps up the reason to prefer KDE over Gnome and the reason to prefer Debian (Ubuntu, actually, and admittedly it makes a huge difference here) over Fedora Core in his recent Free Agent column on PC World.

In Gnome, there's no built-in way to customize your Applications menu (the Start menu equivalent). There used to be a means to do this, but it was removed in a previous version of Gnome, and no new functionality has emerged in its place. I am lucky that the Ubuntu Applications menu is so trim and organized. I have yet to play with Fedora Core 4, but I understand its users are not so lucky.

I didn't realize that apparently everybody in Utah knows by now what silly company SCO is - and by everybody, I mean non-techies who can't program their VCR etc., too.

From Salt Lake City Weekly, a decidedly non-technical publication:

We [Utahonians] believe in the immeasurable societal good that stems from the unequal treatment of gays, in much the same way that we believe in Intelligent Design, cold fusion and The SCO Group’s ability to make money off something other than patent litigation.

Of course it's not a patent litigation, and of course they won't make money from litigation either - but I'm still non-plussed that a juornalist who certainly is not interested in the case would assume his readers to understand this here

An old idea (1993, see at end) redone, and with style: The Optimus keyboard created by the Moscow (did you know, there's no snow in Moscow in summer. Says so in the FAQ on that page.) 'Art. Lebedev Studio': a display on every single key. Should supposedly go in production next year, and there should be an SDK. Hopefully with lots of detailed documentation, so that X11 drivers will be doable.

This is a reply to Network World's open letter to the open source community (taken up only today on TechWorld.)

2005-07-22: Article removed - published on TechWorld.

In an essay in Prospect Magazine, Bill McGuire sketches possible disasters that could befall our world and that society - we - should prepare to deal with global catastrophic events.

Research that allows us to predict earthquaqes, tsunamis, vulcanic events etc. better and allows us to limit casualties by timely evacuation and so forth is certainly money well spent. But I disagree with his view that we should prepare for global events like decades-long cosmic/vulcanic winters and similar things. While it certainly would be technically feasible, I claim that the cost would simply be too high, compared with the risk. Using these resources (money, brainpower, ...) on other problems (poverty, hunger, education, ...) does just get us a better return on investment: Clear benefits can be had now, and can be had with high probability, vs. doubtful benefits (did we really prepare for the right things?) in a few decades or even centuries - if at all.

Now this is not strictly a 'Linux Rulez' item, but an 'Open Source Software Rulez' item: I love the British, and just now, I love the BBC because their way of supporting Free Software makes sense: action speaks loudar than words!

I have seen it claimed (links welcome, I've forgotten where), that the Total Cost of Ownership might not be the right metric to base a proprietary vs. open source (specifically, of course, Microsoft vs. Linux) decision on. In my opinion, TOC is indeed the right thing to measure - but as with all metrics, it's very easy to tweak it to what you'd like it to say. So, the advice is: use TOC as your metric in business use of IT - you're a business, so you gotta make money. But, and this is very important, don't believe the TOC values you see in any study (regardless of who has paid for it.)

So here's a quick guide on what factors to take into account when determining how much your IT infrastructure is costing you: First, running costs: Software licenses (MS world: a lot. Linux world: not much, if any at all.) Hardware replacements (Won't differ a lot between MS and Linux.) Regular software upgrades (Security upgrades: costs might be similar if your staff knows its business. New software versions: Proprietary software companies traditionally stop supporting old software after a time. Open source software is more complicated: you will always be able to find somebody to support any (old) version of your software, if a software upgrade is too complex and expensive. Of course, as users of your obsolete software version become rare, this support will become more and more expensive, and you'll be essentially maintaining a pure in-house software project, losing most of the advantages of open source software.) Data security (Costs for making backups are probably equivalent. Viruses and Trojans are currently an almost non-existing problem in the open source world.) Support (I do have the impression that OSS does give you fewer headaches, but OTOH your users might be less familiar with OSS, mitigating this somewhat.)

Then, since you're comparing Linux vs. Windows TOC, you're probably thinking about switching over. So don't forget the cost of a migration - personnel has to be trained, document templates have to be ported, accessibility to old data has to be assured. Depending on your environment, especially the density of in-house customisations, these costs will vary widely. In any case: time your migration with the 'end of life' of your most important applications - if you would need to upgrade those anyway (Windows 2000 to XP migration, whatever), it's a good point to look into using Linux (and, of course, all the other OSS products.) Platform upgrades do cost, be it Windows - Windows, or be it a Windows - Linux platform change. Obviously, a Windows to Linux migration will be more expensive initially, especially if your IT staff doesn't know Linux well. Just make sure you get some experienced troops in early in the project, and not only when you feel that the crew is at the end of their knowledge. Additionally, don't forget that interoperability is not so bad: you can do the migration step by step. OpenOffice.org will read your MS Office documents in 95% of all cases. And it also runs on Windows - so, perhaps, first migrate from Internet Explorer and Outlook (Express) to Mozilla Firefox and Thunderbird (possibly also from Outlook to a web based replacement like OpenGroupware.org, Kolab or PHPGroupware) and from MS Office to OpenOffice.org. Then, once your users know these applications, a change of the underlying operating system will be quite trivial from a user perspective.

Finally, don't forget the risks you're taking with your IT infrastructure - here is the point where almost all TOC studies fail, because it's very hard to put a price to these things. With proprietary software, your biggest single risk is that your vendor will decide that the product you're using is obsolete and will not get any new updates. This can happen very quickly in the course of company mergers etc. - and there is just no easy way for you as a user to handle this sensibly: even if you're prepared to throw some money at the problem, you'll be forced to migrate to a different product in all likelyhood - and migrations in the proprietary software world are often quite expensive, because customer lock-in is the watchword: making it easy for customers to migrate away from their products is not in the interest of the vendors. The open source world is very different here: vendor lock-in is just not an issue, and as long as there is a userbase, there will be people with an interest in continuing development of your application. But using open source software, you will now likely have to bear all the risks of a software vendor, specifically patent and copyright hassles, because there's no vendor who is taking these for you. To this, there are two answers: one, you're not alone - all users of an open source project share these risks, and once legal wars start, there's a big chance that you'll be able to pool the resources of many users. Two, there are now companies like OSRM who will insure these risks for you. Besides, it's not very likely that you'll ever have to face this sort of problems - as an end-user, you're not generating revenue from distributing the software, so there's not easy stream of revenue that can be claimed as damages. Also, unless you're a very big company, you're not an attractive target for landsharks.

My conclusion in the proprietary vs. open source issue: Basically, assume that the costs of your IT infrastructure will stay the same. Put the money you save by not having to pay license fees to good use in better training for your IT staff, and plan your migrations with care, and you'll not be disappointed. Plan a migration to open source software with the intention of saving some money quickly, and you'll be punished by being the creator of yet another IT project where the budget exploded to the point where you need to print it in landscape paper orientation because the numbers won't fit otherwise.