[Discuss] SQL discussion

[sweetser at alum.mit.edu (Doug)]

postgres rocks in my opinion.  But I still have a fear: triggers.  It is
triggers that blur what I view as a database - place for my data, with
programming, which is to transform the data.  There is no universal trigger
language.  In postgres, there are a bunch.  The trigger code is not so
pretty.  In an ideal world, I would never use triggers in a db, saving that
transformation work for python using pyscopg2.

On Tue, Jan 13, 2015 at 11:10 AM, Mike Small <smallm at panix.com> wrote:

> markw at mohawksoft.com writes:
>
> > I'm a software engineer and I am constantly confounded by other
> engineer's
> > trepidation/apprehension/dislike for the common database. SQL databases
> > especially.
> >
> ...
> > How much of this is a reluctance to learn SQL?
>
> I think the answer is more in Kent's earlier quip about it reminding him
> of Fortran (or others with different backgrounds might say COBOL). I
> resisted for a time, but my smart alec response to that is, "have you
> tried writing all your queries in lower case instead of upper?"  But you
> can't deny there's a definite revulsion from certain quarters despite
> how useful SQL and RDBMSes turned out to be in industry.
>
> Suggesting it's reluctance to learn is just begging the question. The
> people involved are generally more than happy to learn Python, Perl, the
> Go programming language (off topic: is Go a real thing now or just a
> recruiter bait and switch tool?), physics, you name it. Why don't they
> want to learn it, why the revulsion? One programmer I worked with, who
> came without SQL and relational database experience, expressed it with
> frustration: "it seems like this should be really simple and obvious but
> it's not coming to me somehow."  My theory is that it's like discrete
> math as if expressed by an accountant. If it could somehow be expressed
> elegantly (Tutorial D?), these people might like it. Or else if they
> learned relational theory first and then afterwords learned the SQL
> language and whatever RDBMSes they have to, as a means to their ends
> since they're working for money and have to make such compromises, then
> maybe they'd be okay.
>
> How the recent NoSQL popularity fits into this, I'm not sure. Haven't
> been exposed to that yet.
>
> --
> Mike Small
> smallm at panix.com
> _______________________________________________
> Discuss mailing list
> Discuss at blu.org
> http://lists.blu.org/mailman/listinfo/discuss
>