How to get the complete set of history pages for an article.

1. Click on history, and click on 500.
2. In your address bar, change the URL to 5000 by adding a zero, and reload.
   You probably have the whole enchilada now on one page.
3. Save the page as a plain text source file, or download it with curl.
4. Write a script to extract the ID numbers, and insert these
   numbers into a curl command that runs on Linux.

Here's a little BASIC program I used to do this, which gives you an idea
of how the script should work:

OPEN "DBHIST.TXT" FOR INPUT AS #1
OPEN "IDOUT" FOR OUTPUT AS #2
STARTIT:
IF EOF(1) THEN CLOSE : END
LINE INPUT #1, A$
C = INSTR(A$, "oldid=")
IF C = 0 THEN GOTO STARTIT
B$ = MID$(A$, C + 6)
C$ = LEFT$(B$, 9)
' the number could be either 8 or 9 digits; strip last character if necessary
IF RIGHT$(C$, 1) = CHR$(34) THEN D$ = LEFT$(C$, LEN(C$) - 1) ELSE D$ = C$
PRINT #2, "curl -o ";
PRINT #2, D$;
PRINT #2, " -s http://en.wikipedia.org/w/index.php?title=Daniel_Brandt\&oldid=";
PRINT #2, D$
PRINT #2, "sleep 3"
GOTO STARTIT

That program will give you lines that look like this:

curl -o 123211686 -s http://en.wikipedia.org/w/index.php?title=Daniel_Brandt\&oldid=123211686
sleep 3
curl -o 123149927 -s http://en.wikipedia.org/w/index.php?title=Daniel_Brandt\&oldid=123149927
sleep 3
curl -o 122075839 -s http://en.wikipedia.org/w/index.php?title=Daniel_Brandt\&oldid=122075839
sleep 3

Run the script. If you are running curl from Windows, you need quotations marks
around the URL instead of the escape in front of the ampersand.

You end up with a file for each revision, where the filename is the ID number.
If you want to make the file appear the way it did on Wikipedia, you then have
to insert this in the header of each file: <base href="http://en.wikipedia.org/w/">

This line in the header allows you to pick up the extra fluff directly from Wikipedia
when you view the file from your local disk using your browser. The DBHIST.TXT that
you got in step 3 is a handy index to the files (with edit summaries and editor's name
and date stamp). You'll probably only want to print out the juicy ones (there were
2,444 files for my bio page). Be sure to save all the Talk archive pages too; those
are always juicy, and they are as "published" as the article pages because Jimbo lets
them get crawled by the search engines.

Now the bad guys can wipe your history, and you still have it! I wonder if I could start
a business as a forensic archivist for victims of Wikipedia. I could charge outrageous
fees to lawyers!