mmc

"media companies make a huge chunk of their income from buyers who lost their original copies and are forced to buy new ones."

That's new to me - is there a reference somewhere for how big a chunk this is?

I loved some of these - that Cray "Bubbles" picture is classic, and even though I never saw a ZX Spectrum, the color scheme is awesome 80's. Some are a miss, though - the CPC 464?

If I were making a list, I'd add the Connection Machine CM-2: http://www.digibarn.com/collections/parts/connection-machine... Now that's what a supercomputer should look like, not all these boring racks of commodity systems we get nowadays.

There should also be something from SGI - a bunch of pics are here: http://www.geocities.com/sgipics/

I'd also add the Tera MTA, a supercomputer system I'd bet few people have seen: http://www.geocities.com/Athens/6270/su-tera.jpg

When I posted this, there were 4 minutes left until launch

Some great advice. If you treat college as another means to an end, you might become exactly what you wanted to be when you were 17. Interesting people I know all agree that their 17-year-old self couldn't know what their 25 or 30 year-old self wanted.

Therefore, the best path is to get good grades, learn about what interests you and otherwise, just concentrate on becoming your older self well. Sometimes the order of those three is not as important.

No. My experience in a PhD program made me more employable. Last year, when looking for a job, I got interest and offers from industry positions as a direct result of the skills I learned while pursuing a PhD (in Computer Science). As kurtosis here mentions, you often learn real practical skills while doing research. I interviewed for developer positions, and no one suggested that I wasn't practical. We talked about development I'd done.

The article's use of the "plain ol' developer position" phrase suggests that this person had a different problem. God forbid they said that in an interview... They were apparently not interviewing for jobs they actually wanted.

If you actually want the job, you'll be interested in it, you'll explain why you're a good fit, and you probably won't describe it as a "plain ol'" anything, even in jest, even to yourself. Otherwise, why waste the time?

Hopefully I know enough about a co-worker's skills to form a good opinion of them before I find out what college they went to. If you bring it up in our first conversation, I'm going to wonder why.

On the other hand, if you went to grad school, research is such a small world that this kind of thing comes up more often - I might ask who your advisor was, etc, to find out whether we have friends in common. It's less about skills.

On reading the title, I expected this to be about persistent multiplexed sessions in the window server. Adding window-dressing to gnu screen is a strange idea - does the target audience for screen really want this?

"Fast"

I tried it the other day on my macbook pro (2.5Ghz), and it skipped several video frames every minute or so. The audio was fine, but the skipping was jarring and unwatchable.

The previous player ran very well on a five year old Dell laptop.

This is a really interesting idea. A way of signaling quality of publications that is more useful than 'it made it through the committee at some conference/journal/etc' would be a major improvement for researchers.

The graph analysis is even more interesting - this is a way around the problem that publication decisions can be made based on reviews written by less-well-read grad students instead of the faculty one might prefer to hear from.

One question though - it seems like if this guy really wanted to sell the idea, he'd point us to at least one instance of it being used. All it'd take is asking someone to review an existing paper of his (or something). It'd be great to tinker with the review graph analysis, for example.

Why not even share a signed review of the project itself? I tried to generate one for this post, but it doesn't seem to work on OS X, so I filed a bug instead.

Because so many people use it for business communication, I think it'd have long-term effects on the economy that would be really difficult (but interesting) to measure.

It's not a perfect example, because you can get your mails out of GMail, but - This kind of thing might be what it takes to pass legislation forcing data services to provide access (see http://ascii.textfiles.com/archives/1617 )

OK, so in six months, four new mail clients will have conversation view and Apple will add it once those have some success. But how many will be hackable?

In my view, the real problem with email clients (that aren't emacs) is that I have no good way of hacking the filtering, display, and handling of my messages.

I completely forgot about ECB! I don't use it with Python (for no good reason), but it was really great for C++ projects. Love ECB.

I long used emacs in X11, but recently switched to Aquamacs out of sheer laziness - not wanting to recompile emacs+gtk for my new intel mac. I've tried all the various ports, and Aquamacs is my favorite 'native' port, but I actually miss the X11+gtk build I used before. There's something to be said for an emacs that feels like emacs :)

I agree that the option key is no good as Meta - Cmd is Meta in my setup.

I absolutely override the Mac keybindings. I like having M-Q as fill-paragraph, and I found that other useful commands were being hijacked by useless OS X Text Services (like M-> for end of buffer got stolen by OmniOutliner)

I have a large 1920x1200 display that I tend to use with one emacs frame maximized, and some arrangement of windows inside that frame - most often I have a full height window on the left, and two half-height windows on the right. Sometimes I use three columns of windows, and other times four equal-sized windows makes sense.

One of those windows is almost always running zsh inside ansi-term (M-x ansi-term). Some prefer shell-mode, but I think ansi-term and zsh are more useful. I don't think shell-mode and zsh cooperate too well.

Most of my work is done in python, and I just started using pymacs and ropemacs, which seem useful. Notes on them are here: http://www.enigmacurry.com/2008/05/09/emacs-as-a-powerful-py...

This is an interesting article and I think a good overall point to make, however the generalizations don't all hold up.

Based on personal observation as a long-time CS grad student, I take issue with #1 and #2.

Grad school only delays adulthood if you let it, and I know plenty of people who matured well during grad school as well as plenty who did not while working "real" jobs. You can also argue about whether delaying adulthood is actually "pointless" :)

There are many reasons to get a PhD in CS if you are not going to teach - industry research is active in many sub-fields, and those jobs often require a PhD - and if not, it certainly helps.

Finally, in my admittedly small sample, of people I know who graduated with a CS PhD, at least three received six-figure job offers from industry. I'm not sure what adjusting for IQ, education, and working hours looks like, but people don't make decisions adjusted for IQ - they are looking for absolute dollars, and a PhD can deliver.

Again, this is based on CS experience - I have heard less encouraging things for (e.g.) biology.

Some good points - I agree with hs saying 'do the scariest thing'. While you're young. I think this means either the startup or the PhD. But you sound more interested in the startup - you didn't even mention your intended research area!

A PhD is lots of really hard and thankless work, and if you're not really excited about it at the beginning, you won't finish, and that's the worst thing you can do.

So - unless you're hiding your passion for research in your post, I recommend the startup.

I'm sure you know about them, but just in case: the AML field as represented in the Wikipedia article (and the sales report PDF it linked to) missed one company I know of who's been doing AML services 'since 1993', called Logica: http://www.logica.com/risk+and+compliance/350234053

First time I've seen a blocker design for visual confusion, but some existing leg pad designs do incorporate this kind of visual trick to attempt to make the 5-hole look bigger.

Here's a long list of thumbnails of pads: http://www.hockeyworld.com/catHome.ihtml?catID=30&shop=0...

Note that while some just try to look cool, there is a common pattern of diagonal stripes from the inside of the knee to the outside of the foot - when you go down into a butterfly, the idea is that this makes it look like your knees are bent more than they are, and the five-hole is bigger than it is, because the insides of the foot and shin are white like the inside of the crease.

The Vaughan VPG 6* look the most like they're trying to fool the eye. Interestingly, others come close but ruin it by letting a dark pattern repeat right at the inside of the toe.

I thought that this was a Patrick Roy innovation, but couldn't remember where I read about it first.