To deter bots from scamming our addresses and including us in Spam mailing
lists we've tried to make it difficult to analyze & incorporate the
Manually type the basic (un-accented) Latin character equivalents of the text
found below into your Email client's "To:" field. You must do this
since most of the displayed characters are unicode (two byte) values that are
not supported by most Email servers or clients. Attempting to "block and copy"
the text may crash your Email program; it will definitely put garbage into the
Older browser programs will not be able to display the characters properly; they will show a "?" or empty rectangle in the unrecognized character positions in the string. Browsers vintage 2003 and later should recognize and display all of the characters properly ... we've used extended ASCII, Greek, Cyrillic and extended Latin characters to create the displayed string of "letters." Good luck decoding them!
Page last updated on 11/15/2011
|A NOTE TO ALL OF YOU WHO SENT US
EMAIL DURING THE PERIOD APRIL 1, 2004 to SEPTEMBER 1, 2004:|
A misearable convergence of horrible luck reduced our email access and files to binary rubble. Twice the university's massive data storage system suffered catastrophic failures of hard drives across much of the system. The reloads were incomplete and as much as 25% of the email in our account was lost. A month later, it happened again, with another 25% or so getting hosed. What was recoverable was a giagantic raw text file without any indexing or headers --- we obtained it and loaded onto our machine in hopes of salvaging some of it.
Then wave after wave of malware hit our network & we wound up reloading our entire system four times from early June through the end of August (thanks to a 20Gb tape backup system we saved the entire text file + about 90% of what had come in fresh to our mail accounts).
At the time the U had no spam filtering, so every last piece of dog crap with three of our email addresses on it went into the text file --- almost a million messages --- a couple of Gigs to be worked through.
Anyway, as of the end-of-year 2010 we were still pouring through it all, trying to salvage as much as possible (some of has been reduced to just plain gibberish & completely useless while other parts have messages broken up into non-contiguous blocks). What started out as a 2Gb jigsaw puzzle on soggy cardboard pieces and missing picture elements has been reduced to about 75Mb of possibly fixable message text.
So, for those of you that wonder what the heck is going on --- why we didn't answer your (sometimes repeat) messages ... well, they probably just went into the big data mixmaster that the U's servers & our local machine became.
We are spending about 5 hours every other week trying to rebuild a few complete, or at least comprehendable, messages as well as weeding-out the remaining garbage ... maybe in another year we'll have gotten through all of it. We won't be responding to any of it ... too many lost, damaged & pure garbage headers. At best we hope to make corrections to the databases based upon what we are able to pick out in the file.
The three lessons from all of this ---  backing up your system regularly really pays;  don't buy hard drives that are offered-up "on special". Imagine having several dozen hard drives all fail within the course of a single day, just a few days after their warrantees ran out. Then imagine it happening twice on a second set of installed drives about a month later. You say "What's that sound ... like muffled fingernails screaching on a chalkboard? Why, that's the sound of 30 hard drives all experiencing a simultaneous head crash!" And  don't trust your email to servers maintained by low-paid undergraduates working their way through school doing grunt system administrative duties (also in 2004, one of them made a copy of the master user authentication file and sold it to a spammer ... the U had to lock everyone's account until they applied a new password to that account ... all 60,000+ accounts ... what a mess). Our email is now hosted on servers that we pay a monthly fee for ... hopefully they are a bit more motivated to do the job right. So far, no big disasters like we went through in 2004 (knock wood).
UPDATE: As of 1/1/2011 we've given up trying to pour through the remaining cluster-foxtrot of a text file ... too many inscrutible text snippets embeded in pure garbage, not to mention the unicode issues (when the code set byte doesn't make sense).