Personally I'm waiting for them to add better integration of PPAs into Synaptic.

Well unless the authors become dumbasses overnight, you'll probably be waiting a long time. Package management needs to be a single coherent database, making it much more distributed than it needs to be is just asking for pain ... PPAs/KoPeRs aren't terrible in moderation, and solve a couple of problems. But if you make them easily available (ie. available to people who don't know what problems they cause) the solution is much worse than the problem.

Alas. whenever they try and get a new computer the websites don't work in their web browser.

Less sarcastically, assuming there is a correlation between "runs new web browser" and "buys my product" is not the same as having actual data (and I've never seen anyone with actual data make this kind of decision).

First problem is they require bzr 1.16.1 to download their rocket-fuel-setup script, the latest available version in the Ubuntu repo is 1.13.1 -- so you have to manually add the PPA source.

bzr-1.16.1-1.fc11 is the current version in Fedora 11 :)

This is the Unix philosophy: Write programs that do one thing and do it well

Like? "cat -n" or "nl"? make? Awk? Perl? bash/zsh? Which mythical version of unix is this which has programs which only do one thing, but do it well?

The absolute worst kind of cat owner is the ones that let kitty roam the neighborhood free. That is wrong and incredibly rude. Kiltty can die or get hurt in tons of ways [outside]

The absolute worst kind of parent is the one who doesn't lock their child in the basement. That is wrong and incredibly rude. Children can die or get hurt in tons of ways [outside]

You can forego having a real UPS on your live servers too, but that doesn't mean it's a good idea.

You can have all your production servers be z10 mainframes too, doesn't mean it's a good (or cheap) idea.

RAID5 write hole due to system crash (or power loss) between data and parity updates. Resulting in loss of redundancy and eventual data corruption.

It's easy to have pairs of RAID1 drives in a RAID0, no RAID5 no RAID5 write hole.

if your boot drive fails in a manner that allows access to bootsector but blocks access to the kernel image on Drive0, the system will not boo

Why would /boot not contain the kernel you need to boot? This is an automatic default setup if you choose SW RAID in Eg. RHEL.

Array health monitoring does not display red lights on failed drives, as it does on an integrated RAID controller.

Bullshit. In fact it's not too hard to setup soft. RAID setups where when you issue commands to drop a drive the red led above the drive starts flashing, until the tech. replaces it. It's also possible to have hot swap.

Integrated RAID devices typically integrate with system monitoring software and can send proper alerts to admins via SNMP and e-mail, in a manner that integrates with common production grade monitoring solutions. On a system running mdadm, there is no method of doing so, short of cobbling together an ad-hoc script, that would be error prone.

Riiight, madm etc. doesn't integrate with SNMP.

Of course the "HW" RAID is much more expensive, operates like a black box sometimes leaving you totally screwed if the HW dies (esp. any of the cheaper solutions, which I don't think you were advocating but is what people tend to use instead of SW RAID when seeing rants like yours).

I'm not saying it should always be used, sometimes the cost really is worth it, just as it is with PostgreSQL vs. Oracle ... but to dismiss it out of hand like you do is insane, IMNSHO.

I guess that's possible, although I wonder if you included the fact the company invested in knowledge they still have or that Oracle is a serious POS to setup properly (and really needs a dedicated DBA) ... and the fact you'll be paying for the Oracle licenses forever, and that the cost of those licences will dictate what types of HW you can use (want to use a set of 16 "cheap" x86_64 boxes ... sucks to be you).

And while that is possible, personally I find it much more common that a company is paying significant sums of money to Oracle when they could just type "yum install postgresql-server" on their RHEL box.

I can honestly say that the feel and smoothness of the Mac OS X GUI blows x.org out of the water

Here let me fix that for you:

FYI I have yet to hear anyone knowledgeable say that X.org is slow for any reason other than driver quality. HTH. HAND.

You seem to be under the misconception that dynamic linking gets around the GPL. It doesn't. It can, in certain cases, make a difference for the LGPL.

You seem to be under the misconception that the above are facts, and not just the opinions of some lawyers. "The most widely repeated opinion, is that you can't dynamically link against GPL code" is probably the most truthful thing to say.

PolicyKit is an excellent way to handle permissions in a GUI - far better than GtkSudo, etc.

Or not. Try doing an install in PackageKit, and then removing the same thing ... you get asked to authenticate twice (retarded), and if the removal is allowed there's no way to deny that same user from removing glibc as well (double plus retarded). Now it's possible this is all PackageKit is PolicyKit is awesome, but AFAICT that's not true and the usual answer seems to be along the lines of "PolicyKit is designed to let a normal local user change the time from the gnome-panel, and works as long as your problem domain is of a similar level of complexity.".

Also ... please don't think about what happens when you have GUI applications that have arbitrary privilages, ptrace() doesn't exist etc. (actually the std. PolicyKit response is to disallow ptrace()/core/etc. on startup ... and pretend that's clever, and actually works).

I think that's a little bit disingenuous. Try searching "reddit.com" of google and bing. Google gives you a list of all results mentioning reddit.com, and a few common links into the site. That's it.

Err, what? The first link on google is the same as the only result you get on "bing". The wikipedia entry is link 3 on google, not anywhere on bing. Now try searching for "redit.com", first link on google is "maybe you meant reddit.com". All the result on bing look like spammy comercial credit card sites.

Crohn's & Colitis are linked for a reason, in fact the way to work out if you have Colitis is usually to test for Crohn's and if the test is negative you probably do (although you might still have Chrohn's).

Saying that, C&C are "infamous" for being undiagnosed/misdiagnosed.

However the references to that block, then have to be changed (and the references to them, all the way up the tree). But in the real world applications don't write one byte at a time.

By definition. If the OS can't schedule it immediately on 4 cores, but can on 3 cores then it can split into 3 threads. No OS I know of can do this or has an API to do this.

I would assume every OS has an API to do this, but it might not be as fast as you want. Certainly on a modern Linux you can look at columns 6 and 7 in /proc/schedstat for each CPU and get a pretty good idea on CPU idleness. But it's never going to be magic, just as it's never going to be as fast to create a thread as it is to call a function.

I'd also argue that threads are way overused, if you have a specific need and a defined interface ... then create a new program and use shared memory (or whatever) to transfer data. Threads are mostly used due to a lack of design.

I like the new income tax brackets you guys are installing now. [...] So if you earn £150,000 a year [...] The VAT is such a tax on the poor, it's not even funny.

A couple of points: 1) £150,000 is $243,000 dollars atm. ... you can't even see poor from this point. 2) If you include medical and retirement, and add Federal+state taxes, the US taxes are close to UK ones. 3) While I agree that all purchase taxes are regressive, in both the UK and US they at least tend to exempt basic necessities for instance most food if VAT free in the UK (although, again, at 240k+ a year you are probably spending most of your money on non-necessities).