DA Morgan wrote:
> Serge Rielau wrote:
>
>>> Ah yes, I forgot: you have seen the Oracle code and worked
>>> with it. As such you can confirm that indeed DB2's is
>>> better "structured".
>>
>>
>> "That" left some room for interpretation. I know for fact that DB2's
>> code is very well encapsulated w.r.t. OS APIs. I know for fact that
>> DB2's porting team is a lot smaller that Oracle's porting team.
>> If I saw Oracle source code my first call would be to an IBM IP
>> lawyer, my second to Mark Townsend. I sure wouldn't devulge it here.
>>
>> Cheers
>> Serge
>
>
> Given that you have never seen Oracle's source code. You should never
> have commented.
>
> The size of a porting team has little if anything to do with how well
> or poorly anything is written or encapsulated. It may well be that one
> team is larger than another because of the number of products it ports.
> It might be larger because of the complexity of the product ... DB2
> sure doesn't need anyone to port RAC or Grid or Partitions: Capabilities
> it doesn't have.
>
> Then again it might just be that one company has a larger testing
> component associated with the team because they are more concerned with
> quality or creating on-line documentation or any number of other
> possibilities.
I suppose the reasons cannot be confirmed by either of us and Mark T.
silence may be interpreted either way.
If Oracle lumps QA and ID into their porting division that would be a
surprising organization to me though.
When did O10gR2 ship first btw? If I were a Windows shop and had to wait
for sevaral months compared to Linux or Sun I would ask questions.
The reasons you bring forward are about human resource allocation. It
speaks to me about commitment.
A product that uses good encapsulation does not require the same amount
of QA across all combinations of platforms because only low level
function can diverge.
A product that uses good encapsulation does not require extensively
customized documentation per platform. If I need to learn a different
set of docs for a different platform that does not speak well for
portability. The platform differences for DB2 for LUW are so small that
they fit into paragraphs at the appropriate manuals (like setting up a
c-compiler or where digianostics are dumped, ...).
I do not know how well Oracle's codebase is encapsulating the OS APIs,
but your arguments, if true, are not disarming the correlation.
Deep exploitation of OS across major code segments (such as RAC) is very
expensive to support and now we are really not talking same codebase
anymore are we? Which is the haggling point of this subthread.
So what about we call it a draw? :-)
> Your comment was baseless and immitating MarkA's posturing is not
> flattering to your otherwise professional reputation.
Which comment, that I know DB2 source code? Hmm. Well I'm surprised you
have granted me a professional reputation to begin with, so I can live
with the loss.
Cheers
Serge
--
Serge Rielau
DB2 SQL Compiler Development
IBM Toronto Lab
Received on Wed Oct 26 2005 - 16:17:56 CDT
