Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!uxc!uxc.cso.uiuc.edu!gistdev!flint
From: flint@gistdev.UUCP
Newsgroups: comp.unix.wizards
Subject: Re: [ted%NMSU.Edu: ]
Message-ID: <8800021@gistdev>
Date: 14 Jun 89 16:09:00 GMT
References: <19976@adm.BRL.MIL>
Lines: 87
Nf-ID: #R:adm.BRL.MIL:19976:gistdev:8800021:000:1951
Nf-From: gistdev.UUCP!flint    Jun 14 11:09:00 1989


/* Written  5:11 pm  Jun 12, 1989 by nmsu.edu!ted in gistdev:comp.unix.wiza */
/* ---------- "[ted%NMSU.Edu: ]" ---------- */


Guy Harris <guy@auspex.auspex.com> writes

	...
    It would be interesting to see the distribution of file name
    lengths on a BSD system (where the limit is probably essentially
    infinite for all but the most perverse user or application), to
    see if there's a bend in the curve suggesting a lower maximum
    length, and then see how a fixed-length-entry scheme supporting
    that maximum length does vs. the BSD scheme.
	...


i realize it is entirely out of character to provide in a
unix-wizards discussion, but here is the result of just such a survey
made on an active research machine with about 3GB of disk space in
use.  these results were obtained by doing 

	find / -print |sed -e 's/.*\///' |chars-in-line |sort -n |uniq -c

where chars-in-line is a tiny program to count the characters on each
input line.  the work was done as super-user to avoid directory read
problems.

the results were:

  number of names length
	
	      392 1
	     1057 2
	     2694 3
	     4506 4
	     5846 5
	     7003 6
	     8046 7
	     8662 8
	     7036 9
	     6631 10
	     5334 11
	     4432 12
	     2676 13
	     1533 14
	      978 15
	      696 16
	      441 17
	      280 18
	      250 19
	      173 20
	      120 21
	      127 22
	       74 23
	       54 24
	       74 25
	       56 26
	       35 27
	       32 28
	       23 29
	       23 30
	       14 31
		6 32
	       10 33
		5 34
		4 35
		4 36
		3 37
		2 38
		2 39
		1 40
		1 41
		1 42
		1 43
		1 47
		1 51
      
since this is just a static analysis, it won't really tell us
anything about the dynamic profile of file name use.  

the number of very long names is rather surprising (at least to me),
but there is a good indication 255 = infinity as far as file names
are concerned.

hope this helps somebody.
/* End of text from gistdev:comp.unix.wiza */