|where's dag ? email@example.com (1997-01-30)|
|Re: where's dag ? firstname.lastname@example.org (1997-02-02)|
|Re: where's dag ? email@example.com (1997-02-03)|
|Re: where's dag ? firstname.lastname@example.org (1997-02-11)|
|Re: where's dag ? email@example.com (Jerry Pendergraft) (1997-02-20)|
|From:||firstname.lastname@example.org (John Hagerman)|
|Date:||30 Jan 1997 22:32:42 -0500|
I don't get it.
10 years ago or so, some Bright Dudes at AT&T (Gansner, North, and Vo
are documented) wrote the "BellDag" program to generate nice pictures
of DAGs. It was a great tool in the UNIX tradition -- it did a great
job in most cases, it nearly always worked, it was well-featured (it
worked well on cyclic graphs, too!), it ran fast, and it was decently
documented. AT&T didn't officially distribute it, but somehow it got
out and we loved it. (I couldn't have done my PhD without it.) 'dag'
was widely known as the tool to use, why didn't AT&T distribute it?
Then about five years ago, North and Koutsofios put out 'dot', a new
version of 'dag'. 'dot' was more-featured, and you could get it from
AT&T (although I don't remember how). Was 'dot' going to be the
"available" version of 'dag'? I hoped not, it was over-featured and
buggy -- some of my simplest graphs choked it. But it didn't matter,
I had 'dag' and was very happy with it.
Now I've moved, and I want 'dag' again. I found this:
Cool, I thought. But what if it's 'dot'? I sent mail. I got back
the reply (sender name stripped):
I am sorry to inform you that the DAG software is no longer supported
or licensed by Lucent Technologies (formerly AT&T).
(Note: The web page is still up.)
Bummer. What's up? 'dag' was a Great Thing. Why won't AT&T (or
Lucent, or whoever owns it now) distribute it? Yes, I've looked at
daVinci, and I'm not impressed -- it doesn't hold a candle to 'dag'.
'dag' was fine, but is no longer available. 'dot' was over-featured,
and apparently not successful. Imitators tend towards silly features
(color, etc), not solid code. Why is all the simple, good software
disappearing under a mountain of complex, bad junk? I don't get it.
Anybody have a 'dag' binary for an UltraSparc, or even a Sparc20?
Return to the
Search the comp.compilers archives again.