Thoughts on (k)ubuntu

Slutmonkey57b wrote: »

But that doesn't mean that the only choice is "buggy and released" or "nothing at all".

Perfectly true; but the thing is, your buggy and released is my hey-it-just-works. I know this is going to sound flippant and dismissive but the fact is that for a distribution, they want to release the software that people want to see released. When their bug tracker is telling them that, say, few desktop machines and a larger number of laptops are having any problems with a specific package, do they withhold it until they've reached a critical point? Given the heterogeneity of hardware out there, where do you pick the cut-off point?

Of course, this could simply come down to the fact it works on my hardware (and I'd be arguing your side if it didn't ) - but I think it holds on its' own, picking that cut off point is hard, and you'll invariably end up leaving (a few|some|a lot of) people out in the cold. The Debian project for instance, had glacial release cycles for exactly this reason, and ended up frustrating their users. (Then again, they've got about 15,000 packages for each of about 12 architectures, so it's another kettle of fish.)

Slutmonkey57b wrote: »

Therefore a working, closed source, proprietary driver *could* have been provided instead of a broken one.

I don't follow this; how? (I'm predicating this on the 'could' being on the part of the Ubuntu folk, is that right?) This is only possible if the driver provider is bothered to fix and release in a timely manner, and that's not always been the case. So it comes down to my point that it's a hard choice for the distro maintainers - release a driver they can't fix with known problems on some configurations, or omit it and annoy anyone who wants it and may have the correct hardware.

This is changing mind you, with the full release of ATI and Intel specs, we'll see fully open source and more maintainable drivers going into the mainline kernel tree over the next few years; and in fairness, I've found NV at least to be solid and stable, so fair play to them for providing drivers for an unpopular platform. Hopefully they'll join the bandwagon ATI/Intel bandwagon too.

Slutmonkey57b wrote: »

Will linux expect me to create a partition on one of the other drives and assign it a part of the linux file structure before it will install? Given that I can't do that, what are the options?

Ok, that's clearer. Short answer, yes, it will expect space to be given over to it. In theory you could install software anywhere so long as there's a driver that has read/write support, thus it could be possible with ntfs3g. You'd force the packages to be installed to wherever that partition was mounted. Whether you could trust it to the extent you could trust the ext3 or reiser etc drivers, I don't know, I've not used it.

Thing is though, even if you're primary disk is crammed full, and your secondary disk with important data is, say, 75% full, you could repartiton it to a 75%/25% ntfs/ext3fs without losing data - it's certainly possible if you wanted to. Obviously not ideal (and formatting gives me the heebie jeebies), but then under-budgeting for disk space if you're doing anything non-trivial is a losing proposition.

Slutmonkey57b wrote: »

sudo apt-get install module-assistant build-essential fakeroot dh-make debhelper debconf libstdc++5 linux-headers-$(uname -r)

It's funny y'know, I used to be able to rattle out stuff like that, but Ubuntu has domesticated me, I can honestly say I've done little in the way of that over the past couple of years, other than the odd 'apt-cache search foo' 'apt-get install foo'. I mustn't be a Real Linux User.