Apple have quietly discontinued their mail-in repair service in the UK. While the terms of the AppleCare agreement still state that they offer it, and on-site repair in some cases, they have closed their repair centre and are telling customers to take their machines to the nearest authorised service centre. These are few and far between in the UK; for a lot of people the nearest one is several hours away.

Ping Wales have the full story. From the article:

Apparently customers were not the only ones kept in the dark over this:

Last week, at a conference, I met someone called Frits Post. He seemed like a really nice guy, but I couldn't help wondering if he regularly posted here...

I have been told the I am allowed to spend some equipment budget on a nice dual processor system. I am doing work that requires more than 4GB of address space and would benefit from good threading performance. As far as I can see, this means Solaris on Opteron (much as I like SPARC, it doesn't cut it in terms of price / performance for machines under about 8 CPUs these days). Oh, and I need good floating point performance, which rules out Niagara.

Can anyone recommend a good supplier for Opteron machines at a reasonable price? I need a minimum of two CPUs (or cores), a reasonably sized (but not necessarily fast) hard disk, a couple of GB of RAM. It's going to run headless, so no graphics card is required (the Sun workstation line looks nice, but I object to spending money on an nVidia Quadro card that is never going to be used).

Several years after the dot-bomb, I finally realise what all of the fuss was about with regard to Aeron chairs. I am currently visiting The University of Utah to collaborate with a researcher here and today I replaced the default chair I was using with an Aeron chair from an unused office. It is incredibly comfortable, and made even better by the fact that my supervisor doesn't have one.

On the more geeky side, there are a couple of large SGI SMP boxes (one 60 CPU, one 30) for me to play with and a Linux cluster.

Today we warned the technicians that we might need to use a big chunk of disk space. We were told that there might not be enough, since they only had about 200GB free on the RAID array we wanted to use (the one local to the 60 CPU box). Back home, our technicians complained when I dumped a 2GB DVD image in my home directory...

There have been a lot of arguments recently about why OS X sucks for some things, and the blame is often laid at the door of OS X's threading performance. This seemed odd, since OS X gets its threading system from Mach, which was designed to support threading from the ground up. As it turns out, the reason is somewhat more subtle.

OS X sucks at system calls. Due to the Mach+BSD kernel design, system calls are very expensive on OS X. On Mach, the system call cost is around 10x the cost on BSD - on OS X it's some combination of the two depending on what you do. The next thing to realise is that all of the POSIX thread synchronisation mechanisms are implemented using Mach-level synchronisation primitives. This means every time you lock a mutex (for example) you need to dive down through the BSD layer to the Mach layer. This gives you the horrible overhead of checking the permissions on a Mach port (something that sane Microkernels like L4 abandoned), and is expensive. This makes locking operations on OS X much more expensive than on BSD-like kernels. This, in turn, can make threaded code much slower. If you are Adobe, and you are rendering an entire image transformation in a small number of threads, then you will only lock at the beginning and the end of the operations, so it will be nice and fast. If you are doing low-level parallel operations for scientific computing (when not bitching about the OS X kernel), then you should really try harder to persuade someone to buy you a nice Solaris box.

Today, I wrote some code that enumerates, at run-time, all of the classes which conform to a particular protocol (that's implement a given interface, for any Java programmers) and allows the user to select one.

This is why I like Objective-C, and why using languages like C++ feels so painful in comparison. Oh, and for reference, enumerating the classes was around half a dozen lines. Sure, this kind of thing is possible in C++, but only if you build a huge amount of extra stuff on top.

I have recently become interested in the concept of open source hardware. It seems that you can buy FPGAs which provide a reasonable amount of performance for a relatively low cost. This means that you can implement your own CPU etc. designs in Verilog / VHDL. Places like opencores.org allow collaborative work on these designs, and so you can even download someone else's CPU design and use this.

This is all well and good, but FPGAs don't offer the same feature density as even ASICs (although they are re-writable), so this raises the question of whether it would be commercially viable to do a run of ASICs based on an open core, with a motherboard also based on an open design. Would you buy a motherboard / CPU that could run, for example, NetBSD and was entirely open source? How much of a premium would you be willing to pay for such a thing?

I really like the design of the Alpha CPU, and I would be interested in a machine that had a similar core design, and maybe willing to pay around a 50% premium over x86 for the elegance of the system - particularly if I could also download the core design to an FPGA, modify it, and submit changes back for inclusion in the next revision.

One of the real advantages of open source software is that it can easily be compiled for multiple architectures, so once you've written a GCC back-end and a boot loader for your system you suddenly have a huge amount of usable software.

I am rapidly coming to the conclusion that the ideal CPU would have a very simple instruction set, and not even do out-of-order execution. Code running on it should be compiled first into some kind of byte-code, and any re-scheduling should occur in the bytecode JIT - move as much complexity into software as possible, since it's much easier to upgrade (and to configure at run-time).

This time it's my coverage of Linux 2005 making Slashdot and my Project Evil Tutorial on OSNews. Nice - and I'm very glad I'm not the server admin - the new server is running UML, and hasn't been Slashdot-tested yet.

Apple make shiny toys, there is no denying this. It is a shame that their customer support and build quality are so dire. About a year ago, I decided I should send my PowerBook in for repairs. It was not an easy decision, there were a few things wrong with it (white spots on the screen, one of the SO-DIMM slots not working, detection of headphones had an intermittent fault), but nothing that made it unusable - and I used it a lot. Never mind, I thought, it usually only takes them two days to do a repair, so I'll back everything up, grit my teeth, and send it in.

A month later they finally admitted that they'd lost it. A month after that they got around to replacing it. The one they replaced it with didn't even boot. They collected it and fixed it relatively quickly, leaving me with a new machine which had a fault in one of the SO-DIMM slots.

At this point, I decided not to send it back in for repair again. I'd already wasted around 10 hours on hold to Apple's customer support, and I didn't want to spend any longer.

Now, however, the fault with one slot has become an intermittent fault with both. Moving the machine at all causes memory errors (and will cause it not to resume from standby). When I boot, I have no idea which RAM slot will be operational (I've put a stick in each - sometimes I boot with 512MB, sometimes with 256MB. So far, I haven't managed to boot with 768MB, but there's hope...).

I am going to try to borrow a machine from the department while I wait for this one to be repaired. If I am really lucky, I'll be able to bring it home and work from here. Well, I can hope...

This week I got some new toys. I think my supervisor is feeling guilty (been seeing other students, or something), and practically forced me to spend some grant money on shiny things. These were:

• An Apple 23" Cinema HD. Very nice. I'm using it now. I have so much screen space that I can actually see my desktop in several places (unlike my physical desktop, which is covered with things).
• A LaCie 500GB disk (which is really 466GB). FireWire 800, with FW400 and USB2 as well (although you can only chain them with FW800). I have a 320GB (300GB) one at home, which I use for video editing. This one is for storing large data sets, and definitely not for video editing, oh no. Honest.
• An iSight. Officially so I can join video conferencing meetings with the other 3 sites we work with on my laptop.
• A BlueTooth Keyboard (Apple) and mouse (Kensington - 3 buttons!) so I don't need to plug in lots of wires when I plug in my laptop.

I suddenly have an incentive to stay at my desk (apart from the view of the sea).

In unrelated news, I mentioned previously that I was putting OpenBSD on a co-located Mac Mini, and would be writing about this. The first 4 of these are now online:

Macs in a rack - taking the mini to the masses
Why OpenBSD?
Setting up secure mail on OpenBSD
Filtering Spam with OpenBSD

The fifth one (Web and Webmail) is done, but will not be release until next Friday (it's my birthday next week, so I wrote it early). The sixth will cover setting up a Jabber server. I haven't yet decided what the seventh will cover - suggestions welcome...

Oh, and the hosting place has agreed to give a discount to Ping Wales readers, so don't forget to mention Ping to them if you do decide to use them (details in the first article).

I finally decided to order a Mac Mini. Not as a desktop, but as a colocated server (with the company featured on Slashdot a while back). One thing that they don't advertise is that their `remote hands' service allows you to change operating systems. While I am very pleased with OS X as a desktop, I would rather have something a little more command line oriented on a server (yes, I know you can do most things on the command line with OS X, but it is not really designed for it). To this end, I shall be installing OpenBSD on it as soon as it arrives. This means that I will have a 1.42GHz RISC system running OpenBSD with 20GB of bandwidth a month for $22.99. The closes competitors I found for this were several times the price for a similar service - largely due to the fact that they required 1U systems which cost a whole lot more to host.

Once the system is set up and running, I will post more details.

On my /. front page, I have the two sections I most frequently read in little boxed, BSD and Apple. The Apple box is completely full of April Fools' `Jokes', while the BSD box is completely full of real stories. The seems slightly bizarre, since there are so many opportunities to generate convincing April Fools' stories for *BSD (I seem to recall one last year claiming that FreeBSD was going to start using the Linux kernel), and yet no one bothered. I'll have to remember to write one next year.

Yesterday, Slashdot linked to something I wrote for the first time. Depressingly, this is possibly the least coherent rambling I have ever written (I'd drunk a lot of caffeine and was in a hurry) - even including most of my comments on Slashdot. Ho hum.

Today, I was testing my Jabber client (which now has a really buggy Alpha build released) and I realised that none of my MSN contacts were visible. A short examination of this problem revealed that the entire MSN network seems to have gone down for the second time in 12 hours. Nice to see such good reliability from MS.

I would feel more smug, if not for the fact that my Jabber server (theravensnest.org) was unavailable for a few hours last night due to the fact that the idiots who run the upstream router seem unable to keep it connected for more than a few days at a time. Fortunately, this only affected about 2 people, unlike the MSN glitch.

In other news, my client is now in a semi-usable state, and I am finally able to ditch Psi (which has a habit of deciding to eat all of my CPU after it's been running for a few minutes, probably due to the pile of bugware known as Qt for Mac). In the near future I will begin porting it to GNUStep, which should just require translating the OS X nib files into GNUStep nibs (nibs are serialised object files used for user interface definitions on NeXTStep-derived platforms).

On Monday, Apple announced a whole load of shiny things. Potentially the most interesting of these (and the one that got the least coverage) was the iChat server, which will be part of OS X Server 10.4. It appears to be the Jabberd (or possibly Jabberd2) server, with a configuration GUI. The first reason this is nice is that it means that I can cross writing a Cocoa Jabberd configuration interface and creating a .app bundle off my To-Do list. The second is that iChat will fully support Jabber.

As you may be aware, iChat has used a slightly modified (serverless) version of Jabber for peer to peer chats. Now, it will be using Jabber for through-server traffic as well. At present, the marketing material refers to the fact that iChat server can be used from Windows or Linux using third-party clients, but it would be quite unlike Apple to release something like this without client support on the Windows side. This is especially true if they wish to include support for things like voice and video chats (currently supported by iChat, but not other Jabber clients). While I don't want to make any concrete predictions, it seems quite possible that Apple are going to release a version of iChat for Windows, supporting Jabber. The question then becomes whether the Windows version would support AIM. It seems unlikely that it would, given how much AOL dislike third party clients on their network.

Personally, I'm hoping that Apple will develop a 100% functional AIM <-> Jabber transport for their server, and migrate everything to Jabber (and I said something along these lines on the JDEV list shortly after the launch of iChat). Thoughts?