Re: [pgsql-advocacy] Assembling "top features" list for betaannouncement - Mailing list pgsql-advocacy

Michael, Robert (at bottom):

On 04/08/2017 01:50 AM, Michael Banck wrote:
> Hi,
>
> I think besides assembling the list, it is also important to come up
> with clear, correct and catchy names for each feature.

Thank you so much for the marketing-friendly terminology!

> Am Freitag, den 07.04.2017, 17:40 -0700 schrieb Josh Berkus:
>> So, lemme update the running list:
>>
>> Banner Features:
>>
>> - pglogical
>
> Again, pglogical is a specific out-of-core extension, which is not in
> V10. "Logical Replication" is the one we should promote.

Looks like we can use a more concise version of Robert's explanation for
this.

>
>> - new partitioning
>
> I'd call it "Native Partitioning" or "Declarative Partitioning"

"Declarative", I think

>
>> - quorum sync rep
>
> "Qurom-Based Syncronous Replication"
>
>> - parallel index scan, bitmap scan, and merge join
>
> "Improved and Extended Parallel Query"
>
> IMO the three headline features are Logical Replication, Partitioning
> and Parallel Query - not sure Quorum Commit stands out compared to the
> others below.

I'd argue that Quorum Sync Rep combined with the client connection
failover means that you can easily build synchronous PostgreSQL clusters
out of the box, something which has previously required a lot of
3rd-party tooling.  So it makes my list for major, especially if we
combine the two features.

Given that scale-out has been a weak story for Postgres in the past, I'm
eager to trumpet any and all features we add in this direction.

>
>> Other Features:
>>
>> - XMLTABLE query expressions
>> - hash indexes in wal
>
> "Improved and Crash-Proof Hash Indexes", or "Improved and
> Production-Ready Hash indexes" - there have been several performance
> optimizations as well.
>
>> - PGFDW: push down aggregates
>
> As the list is getting rather large, I am not sure this one merits
> inclusion, but I don't how much it improves things, really.

It's really "more progress in DB federation".  Not momentous, but more
significant than XMLTABLE, which personally I can't even explain.

>> - SCRAM authentication
>
> I'm afraid this doesn't tell a lot of people what's going on, but I
> haven't come up with a better 3-5 word explanation, yet.

"More secure password authentication using SCRAM"

>
>>
>> ... seems like the "other features" list should be longer.
>
> There is also:
>
> "Client-Side Connection-Failover"

See above.

> "Improved and Easier Replication Setup" (Peter mentioned that as well)

Oooh, anyone have a link to details?

>
> Plus what else Peter mentioned.
>
> From a marketing standpoint, I guess it is good that some features like
> WARM did not make it or else we'd have even more trouble coming up with
> a concise list.

Well, it's fine for the beta announcement to have an extended feature
list, although at a certain point we just need to link in a What's New
page on the wiki.

On 04/08/2017 06:39 AM, Robert Haas wrote:
> On Fri, Apr 7, 2017 at 12:39 PM, Josh Berkus <josh@berkus.org> wrote:
>> While 10 isn't 100% baked yet, I'd like to get started sorting out what
>> our "top 2-4" features for the beta announcement are going to be.
>
> I have been thinking about this topic as well, and wrote a blog post
about it.
>
>
http://rhaas.blogspot.com/2017/04/new-features-coming-in-postgresql-10.html
>
> I wasn't able to pare the major features list down to the 2-4 range; I
> found it difficult enough to get it down to 7, and there's still a lot
> of other stuff that is also *really* good.

It's always nice to see that the reason the reviewer pipeline is so
clogged is that we are getting so many contributions.

From a PR standpoint, 5 is the absolute max, and 3 is better.  Really,
the ideal situation is a set of features which can be closely tied
together around a simple theme.

--
Josh Berkus
Containers & Databases Oh My!