Why don't vendors update std utilities?

Why don't vendors update std utilities?

Post by Larry W. Vird » Sun, 28 Jun 1992 02:42:02



Is this standard across all vendors, or have I just had an unlucky string
of encounters?

I have used Unix from at least 6 different vendors.  When I find a problem,
such as sort truncating long lines, or uuencode not protecting against
trailing blanks, man not using MANPATH etc. I try to report the problem.
But then I watch release after release come out with the same problems.

Now in most cases, the software these folks are using came from BSD or
ATT.  But in the case of the software coming from BSD, BSD has released
newer versions which fixed a number of the problems.  The same goes for
ATT.  So why would a vendor continue shipping old code?  Is it simply because
it is easier to do it and ignore customer complaints than to go out and
get the new version and use it?

Obviously, if the vendor has spent a lot of time and effort updating the
source, then going out and getting a new version from the original
vendor is not an option.  And if the original folks haven't corrected
bugs, then I can ALMOST see the reason not to upgrade in that case.

But in at least two vendor's case, it SEEMS like they are using pretty
vanilla BSD code (looking at the strings in the binaries, etc.).  So
it seems like updating to later utilities would be easy enough to do.

Just curious if anyone other than me has encountered this?  The reason it
comes up recently was that I was using a mail interface that allowed me
to attach uuencoded files to mail msgs.  But the uuencode being used was
depending on trailing blanks to be present - when they got dropped somewhere
between myself and the person I was mailing, the file ended up 'corrupt'.
While that is a fairly easy one to fix, one wonders about things like
the BITNET character mangling that goes on with {} and so forth...
--
Larry W. Virden                 UUCP: osu-cis!chemabs!lvirden

Personal: 674 Falls Place,   Reynoldsburg, OH 43068-1614

 
 
 

Why don't vendors update std utilities?

Post by Martin Weitz » Sun, 28 Jun 1992 21:49:26



Quote:

>Is this standard across all vendors, or have I just had an unlucky string
>of encounters?

No, my experiences are similar.

Quote:>I have used Unix from at least 6 different vendors.  When I find a problem,
>such as sort truncating long lines, or uuencode not protecting against
>trailing blanks, man not using MANPATH etc. I try to report the problem.
>But then I watch release after release come out with the same problems.

Probably a lot of vendors seem to care more about restricting their
customers with nonsense copyright notices(%) than correcting bugs.  And it's
still the situation that customers buy software despite known bugs and even
sign copyright licenses that promise NO MORE than replacing the media(!)
if the product doesn't work as advertised.

I see a solution for this, though it requires to convince the courts in
some precedent cases.  It should be considered the right of a customer who
has bought some software to have a product that is free of flaws, at least
such flaws that are well known and corrected.  Considering that in many
cases patches that correct the bugs in the source are available (e.g. on
the net here), it should become legal (i.e. not considered as copyright
infringement), to distribute the whole source of programs for the purpose
of fixing bugs.

I think the following scheme would be a reasonable:  If there is a
software manufacturer M, a system house S who has bought rights on
the source from M and a customer C who doesn't have the right to have
the source.  The customer complains to S because of a bug and supplies
a fix.  If S does not apply the fix and delivers a corrected program
within a reasonable time frame (a few weeks),  C advances to M to inform
about the situation.  M now has the possibility to put pressure onto S
(e.g. he could include some paragraph in its contracts with S so that S
is required to apply the fix on demand of M).

Finally, if neither M nor S care to satisfy the needs of C, any third
(or really fourth) party who also has access to the sources should be
allowed to hand it to C for the purpose of fixing the bug.

I'll give another example that will exactly show what I mean:  Given
you have bought a car and signed a contract with your dealer not to
disassemble certain parts of it.  Now, the car shows some unreliable
behavior (maybe the brakes sometimes do not work) and somebody knows
how to repair it but it requires that you disassemble those parts you
aren't allowed to and your dealer isn't willing to do it for you (free
of charge, of course).  You simply must be allowed to fix it yourself
or let someone else do it, even if it violates the contract.  Point.

----
%: My favorite example is the copyright notice in front of the command
/bin/true, an (otherwise) empty(!) shell script.
--


 
 
 

Why don't vendors update std utilities?

Post by Elizabeth Zwic » Thu, 02 Jul 1992 04:48:30


One of the projects that has been suggested for the SAGE electronic
information distribution working group is the development of a
third-party bugs database. One of my favorite visions of using such a
database involves taking one of these common bugs in a standard
utility, calling up one vendor, and saying "This bug has been reported
against four vendors. One of them issued a patch. One of them said it
would be fixed in the next release. Now it's your turn; you can do as
well as your competitors, or not, and the direct comparison will be in
the database for all the world to see." (It's right up there with "No,
I don't think I am the only person with the bug; right here I see it
being reported by 18 people in 5 different countries.")

More information about joining this working group is available by

System Administrators' Guild, a USENIX special technical group devoted
to system administration interests; to join its mailing list, send

        Elizabeth D. Zwicky

 
 
 

1. WHY-WHY-WHY.....my Mitsumi CD don't seek at boot??!!!!!

....without word...!!!
Is the 5 time that i recompile the kernel with ALL the necessary drivers
for the ATAPI CD, but that stupid Mitsumi doesn't ear to work.
A brief history....the first time i installed Linux the CD, after the
first bootstrap doesn't work, although the CD setup worked fine.
So i reinstalled all (LILO too), and ,i don't now why, this time the CD
worked fine!! BUT, after compiling the Kernel, as above specified, the CD
doesn't work.....('the device appear confuse', and other line).
Many many thanx.....in advance!!

Luciano.....
Rome, Italy. (Not only 'spaghetti'....'mandolino' too!!!!!)

2. Need pcnfsd for Solaris 1.1 on Sun 670

3. Why is OpenBSD's locate update beating FreeBSD's locate update speed?

4. Console Monitor Program for HP

5. Why doesn't 'cut' utility work intuitively?

6. Creative Soundblaster PCI128

7. Why don't 'kill' references appear in pacct/acctcom listings?

8. RESOURCE ALLOCATION AND DEADLOCK AVOIDANCE

9. Why don't console message appear in my ``console'' window?

10. panel crahses at start-up and online update's don't run

11. I don't understand why this code isn't working

12. Why don't you have to tune the linux kernel (like SCO's)

13. I'm a Newbie Baby, So Why Don't Ya Kill Me