[erlang-questions] gen_tcp question

[Sean Hinde]

Joe,

I was not disagreeing with you that it should be documented. I  
suppose I was making an observation on human (and corporate) behaviour.

Personally I am happy to derive an intent from the combination of the  
existing documentation and my own testing. If it ended up not working  
as implied by the documentation then I would have had no hesitation  
to raise a trouble ticket using the commercial support option and  
blag it. Others are not happy to rely on this level of assumption so  
it should of course be documented.

To the point in hand:

I am sure Tony could make some statement of what it was supposed to  
do when he wrote it, but that is surely not enough.

Ericsson might have a plan to change the undocumented behaviour so it  
only works for packets without an F in the length in an unspecified  
number base.

So, only Ericsson can make any statement of what the program is  
supposed to do now and in the future. And that place is the  
documentation and any support contracts you might have with them.

BR,
Sean

On 14 Sep 2006, at 13:53, Joe Armstrong ((TN/EAB)) wrote:

>
> I'm not interested in what is actually does -
> I'm interested in what it's supposed to do.
>
> The documentation is supposed to tell you what the program is supposed
> to do
> and I suppose that it is reasonable to suppose that people are  
> supposed
> to
> send error messages if they suppose that the program does not
> do what might be supposed from the documentation.
>
> <I'm almost proud of that sentence>
>
> /Joe
>
>
>> -----Original Message-----
>> From: Sean Hinde [mailto:sean.hinde@REDACTED]
>> Sent: den 14 september 2006 14:46
>> To: Joe Armstrong (TN/EAB)
>> Cc: erlang-questions@REDACTED; tony@REDACTED
>> Subject: Re: [erlang-questions] gen_tcp question
>>
>>
>> On 14 Sep 2006, at 13:33, Joe Armstrong ((TN/EAB)) wrote:
>>
>>>
>>> It might well be that it
>>>
>>>    ... "seemed to do the right thing" ...
>>
>> Well, It did the right thing consistently for some years with
>> messages much larger that can be carried in a single LAN packet.
>> Personally I would take that as greater proof than reading it
>> in the documentation.
>>
>> Others of a different persuasion would believe that if it was
>> not written in the documentation then I had in fact created a
>> timebomb guaranteed to fail at the worst possible moment in
>> the future !
>>
>> Only Ericsson can make any documented guarantees about the
>> default system, not even Tony can provide anything other that
>> a statement of his intent when he wrote it.
>>
>>
>> Sean
>>
>>