On 23 October 2012 10:46, Simon IJskes - QCG <[email protected]> wrote:

> On 23-10-12 11:17, Dan Creswell wrote:
>
>> Popular isn't good or of value nor does it justify bad design. Apple have
>> stuck with that for a long time before getting to where they are now and
>> they're doing alright.
>>
>
> No sense in having a 'good' design, and not beeing popular. Can you
> predict a timespan where we can see an increase in uptake of river?
>

No sense? Really? Well s**t Apple didn't make any money at all until they
were popular ;)

No I can't foresee River ever being popular because it's different and
requires a different mindset. I can see popularity as the result of bowing
to common practice. Of course, once you've done that, River is just the
same as everything else.

My own preference: Build something good for the right mindset because that
mindset is under-served. It ain't the biggest population but it's a niche
no one owns. And support that mindset you might allow them to build some
systems that bring others into that mindset.

The alternative is to build something that competes in the mainstream with
everything else. And you know what? I don't believe for a second that'll be
made to stick. In that part of the population, River will always be too
complex, too big, too unwieldy. It's too clever for its own good. The world
you're talking of is implicit transactions with annotations, remote calls
that don't fail (cos the network isn't brittle) etc. River is
over-engineered for that and will always be beaten out by other "simpler"
techs. Think you'll succeed? River has a bit of history there already...

You want more examples of this tradeoff? Check out Nosql versus Sql and hey
Basho don't look like they do too bad do they?


> Shall we create two releases of river? The pragmatic one and the purist
> one? I will work on the pragmatic one, and if you see a valuable patch, you
> can merge it in the purist one.
>

It ain't pragmatic versus purist...


>
> Gr. Simon
>
>
>

Reply via email to