[csw-users] Why support Solaris 8 onwards?

Donal McMullan dmcmullan at mac.com
Wed Sep 19 19:05:20 CEST 2007


On 20 Sep 2007, at 04:38, Ben Taylor wrote:

> Donal McMullan wrote:
>
>> I think it was Joel Spolsky who pointed out that code as it gets old,
>> doesn't rust. In fact, the longer it runs without incident, the more
>> trusted and valued it becomes.
>>
>>
> And if you've ever been in the situation trying to get support for an
> obsolete OS,
> you'll recognize that the first time you have to go through that, you
> wonder why
> you didn't upgrade earlier.
>
>> If you want to spend your time replacing battle hardened, proven
>> solutions with the shiniest new thing (in other words, fixing things
>> that aren't broken) it's probably because you're too young to
>> appreciate how precious a commodity time is.
>>
>>
> That's right, I couldn't possibly understand having worked in the
> financial markets 15 years
> ago.  I do recognize, having had the job of coordinating all the open
> bugs with Sun as the
> customer, and doing patch management, the hassles involved in getting
> Sun to support
> an aging OS and getting problem resolution done.   Due to the large  
> Sun
> presence and
> huge support costs paid by this org, several times during my stint,  
> Sun
> had to find
> engineers to come out on site and debug the problem because they
> couldn't reproduce
> it internally, and the guy who came out was typically the person who
> wrote the driver
> or module.
>
> Now, after weeks of limited usability of the systems in question, the
> cost to the financial
> org was pretty high because they had stayed on very old HW and an old
> OS.  So I find
> your justification just as *charming*.  I never want to be in the
> position of having to
> answer why I just cost my organization 50K in lost productivity.
>
> Ben
> _______________________________________________
> users mailing list
> users at lists.blastwave.org
> https://lists.blastwave.org/mailman/listinfo/users


I read a story recently about a Solaris 10 deployment that went wrong
and tanked a project. That doesn't make Solaris 10 a poor OS. It's
just anecdotal evidence - like your story.

You'll note that "if it ain't broke, don't fix it", still leaves you
perfectly free to fix things that *are* 'broke'; in fact I
consider the fixing of broken things to be commendable.

Donal



More information about the users mailing list