Re: Help me improve the 9.2 release announcement! - Mailing list pgsql-advocacy

> "Joshua D. Drake" <jd@commandprompt.com> wrote:
>> On 08/07/2012 02:00 PM, Rob Napier wrote:
>>
>>>> Can someone suggest a way to work range types into a simple,
>>>> 2-part theme which doesn't obscure the JSON message?
>>>
>>> As for the WeirdNewTypes, it needs a simple lo-tech way of
>>> explaining it.  Maybe a short, simple example of how it is
>>> applied.
>>
>> "CoolNewTypes"
>>
>> Weird implies well wierd, they aren't weird, they are cool, hot,
>> nifty, rocking.
>
> Yeah, if we need to punch just two points, with the ability to
> elaborate a little, I would say performance (with emphasis on being
> able to scale well to larger volumes and big new hardware) and new
> types to simplify life for application programmers (JSON and
> ranges).
>
> JSON is all the rage with web-oriented programmers.  There is a
> group of people for whom there pretty much couldn't be bigger
> PostgreSQL news than this feature.  When we get to the day that we
> have an HTTP access method that provides a "RESTful" interface to
> database data through JSON, the web programmmers here will probably
> throw a big party over how easy we've made their work.  9.2 doesn't
> take it all the way there, but it is a giant leap in that direction,
> which will cause some jaws to drop around here.

From a marketing perspective, touting the JSON support will definitely get some more eyeballs from the web community.
Beingin the web in the  "web-oriented programmer" category myself, if I were to further look into the 9.2 JSON support,
Iwould come away wondering what the "hype" is with it as, on a functional basis, it just ensures that I store valid
JSONin the database.  If I had to do any more serious data processing on evaluation and would prefer to do it at the
databaselevel, I would look at another data source. 

As I said before, I'm glad Postgres is starting to have JSON support (I've randomly bugged people for it through the
years:-), but we do need to make sure we point out our strengths. 

> On the other hand, I suspect that ranges will get wider adoption,
> because there are so many places that they make existing queries
> simpler and faster, regardless of whether you are web-oriented.  The
> obvious applications are related to scheduling, but I suspect it
> will be put to many other uses.

+1 and especially for a  "web-oriented developer, " once you understand what ranges can do, you suddenly have a lot of
veryfast querying options :-) For instance, just about anything involving a "slider" UI element that on a web page that
goesinto a search can be turned into a range query.  At the end of the day, most web-developers care about speed and
simpleways to extract performance out of the tools they are using. 

Jonathan