Umm, maildirs store each message in its own file. I clean up (archive) emails from each past year in a separate folder and still easily have 8k files in each... and that is not my busiest mailbox.

After a few thousand items of anything, the proper tool for the job is a database, not a file system. Though file system can be described as a kind of database, any in case there are problems common to both, such as fragmentation, a specialized data storage always beats generic ones. Personally, I like what Dovecot does - maintains a mbox-like structure ("old-fashined", all messages from a single mail folder in a single file) which is also padded appropriately so fields can be updated without rewriting the file) and builds an index file on top of it to enable efficient random message access. In this way you get efficient, big, append-only data files, and small, easily cacheable index files: win-win.

This is Randall's Magnum Opus.

Of course, the real insight is Reverse Phrenology! Kicking people in the head to make them better people!

...except if you have external factors such as patents which effectively prohibit anyone truly new entering the industry ever again...

The OLPC project was always one step near the infamous "Bibles for Haiti" project - a condescending view that an "easy answer", one which is easily mass-manufactured will miraculously solve a hard social problem. That the OLPC-ers are technocratic instead of theocratic makes little difference with regards to the efficiency of the approach. What *should* have been sent are *teachers*, but it's much, much harder to send teachers into the wilderness when they are already so lowly regarded in the western world.

But of course, "eegrep" isn't :)

(enhanced enhaced grep)

"western US" should be "eastern US" of course :)

Hmmm, you do realize that outside the western US, Germany, Japan and some other uptight countries, this is actually the "normal" way of life for the largest part of the world? Some call it "quality of life".

This may or may not be true. In France, people drink wine almost like water. Germany and Belgium are known for their beer. Scandinavian countries' weekeend passtimes is drinking any alcohol they can get their hands on, so would you call all of them "drunks"? Again, this labeling thing mostly seems the problem with the uptight and stressed out USians - the rest of the world works just fine as it is.

Yes, with backup copies on the Google, IBM and Dell clouds in case Amazon goes down (I'm not making this up: it happens).

I concur - they have some amazing things done in their research division but they seem to follow the footsteps of the famed Xerox Palo Alto center - they can't seem to build products on top of that research.

I'm very open source biased but MSR is one of the places I wouldn't have any problems working in.

Hmmm... it looks like the conspiracy theorists should review the Pentagon 9/11 crash :)

In short, RIM was a one trick pony and refused to change, just like Nokia. They did it to themselves.

Unfortunately for the peasants and the third-worlders, there are some huge technological prereqisites:

• You need clean water to wash hands and wounds with - the majority of surface water in "black africa" is contaminated - not by Evil Western Chenicals but by feces and germs
• Covering your mouth when caughing is well and good but to have any resemblance of general care and isolation (i.e. hospitals) you need something to cover your mouth *with*, ranging from clean cloth (see previous issue) to gauzes, bandages and sterile equipment
• Animals in Europe were in houses often for very simple reasons: a) they are warm (remember, the "warm Europe" trend basically started with the 20th century) and b) that was the only option to keep them away from thieves

Basically, I agree with you, but want to emphasize that the ideas need infrastructure.

To be pedantic, that's 65535 per (client_ip,port) pair...

Technically, MS *did* use a valid and acceptedly secure hash functions, DES and MD4. The problem is that, because of backwards compatibility across their 20-year product spans, they were not as vigilant in updating the protocols. Even when they *did* upgrade them, they went to MD5 (with NTLMv2) - which was again proced weak - but they continued to use the older protocol which allowed trivial attacks.

Which is why anyone "worth his salt" will laugh if you propose a crypto system which is supposed to last 20 years and is not flexible in its choice of component algorithms.