Re: MongoDB is Web Scale

On 25/10/2012 4:04 PM, James K. Lowden wrote:
> On Thu, 25 Oct 2012 07:55:25 -0700 (PDT)
> Jan Hidders<hidders_at_gmail.com> wrote:
>
>>> I did actually, and while it strove for a kind of balance, it was
>>> counterproductive in that it falsely implied there was a debate
>>> even to be had.
>>
>> I strongly disagree. There *is* a discussion to be had here and it is
>> raging right now in both academia and industry.
>
> In academia? Et tu, Brute?
>
> I am curious to know: What do you consider the very strongest argument
> by the no-SQL folks? I have never seen one that didn't exhibit perfect
> innocence of reational theory.
>
>> Unless the relationalists are going to realize this and manage to
>> contain their usual arrogance and become more realistic and careful
>> about the claims of the RM they will loose this debate very fast.
>
> Well. The relationalists have been losing the argument for 30 years.
> Ignorance outpaces knowledge in every generation, and the ratio of
> programmers to DBAs is at least 10:1, assuring who'll win the voice
> vote.
>
> The only thing that's kept the RM afloat in this stormy sea is that its
> sturdy construction allows it to continue to work, no matter how abused
> or perverted.
>
> --jkl

The RM ( whether Codd's, Date's, Darwen's, Pascal's, McGoveran's or some other authors whose names I can't remember) has for many years been over-complicated by many of the same people who now push "nosql", probably misunderstood too by most of them, with an industrial theory that is yet not fully formed.

This seems very ironic to me because the nosql movement seems to be very concerned with simplicity of implementation. The logical simplicity of the RM as I understand it remains not well understood at all and some of the famous popular thinkers and writers acknowledge this every so often.   They also get pilloried, wrongly, from time to time for miss-steps. I suspect the difficulty of explaining it to people with an existing computer science heritage makes it seem bigger than it really is. Whereas I imagine that 10 year-old children without that baggage could learn it easier if it were only shown to them, minus cosmetics such as html. The right comparison is hardly between programmers and dba's. The cruel fact is that people of both those ilks will be known far in the future as a lost generation.

And, as is common, here we have 'sql' being conflated with 'relational'.   I think you are right about the 'sturdy construction'. I'd like to see a coherent description of a 'logical nosql model'. Received on Fri Oct 26 2012 - 06:21:25 CEST