Re: The Business Case for PostgreSQL

On Sunday 17 June 2007 23:10, Chris Browne wrote:
> One bit that I think I vigorously disagree with:
>
> "In the same way, adding features you want, whether ordinary ones like
> replication or exotic ones such as complex statistics, is easy."
>
> As a mathematician (couple of Math degrees in my past ;-); I'll put
> this hat on for a moment...), who works on replication, I think this
> is totally backwards.
>
> Implementing complex statistical functions tends to be, behind the
> scenes, a pretty Simple Matter Of Programming.  The formulae may look
> nasty, but the computations are usually not all that bad, once
> understood.
>
> In contrast, I have to call replication a "pretty exotic" feature.
> And I don't think implementing replication is particularly easy,
> either.  The challenges with multimaster are legion...
>

replication is ordinary in the sense that every database system has, and many 
people roll your own.  And I'll note rolling your own in postgresql is quite 
a bit easier than rolling your own in say, sql server (We've done both at 
OmniTI).  My engineering hat tells me that the statement should be tempered 
with a "relativly easy" maybe... or replace easy with "PostgreSQL does a good 
job staying out of your way"... but otherwise I don't it is that far off the 
mark. 

> "From the outset, PostgreSQL was constructed to meet the goals of
> active businesses which could rely on it as a core element of their
> mission-critical IT infrastructure."
>
> Nope.  At the outset, it was constructed as a research project.  When
> it became an OSS project, I'm not sure those were yet the goals.
>

Yeah, this one is probably more problematic.  Could it be argued that, having 
created one system and seeing how it fared commercially, that Stonebreaker & 
Co. did have in mind the idea of POSTGRES becoming a basis for a commercial 
system as well?  Maybe someone from the "Old School" can comment? 

-- 
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL