Re: First Step to Major Use: Integrated Full-Text

Hi there,

All good points, however I want to point out this isn't my personal
issue, since I host my own websites. My only problem is that if I
license my software out to people, and they have the option of using
Postgres or MySQL, they're going to support MySQL because more hosts
support MySQL than those who support Postgres WITH a search module
enabled. I don't have a problem installing Tsearch2 myself, but I'm not
going to responsible for doing it for every single one of my licensees.

I'll break my argument FOR full-text integration in the following few
points:

1. Full-text search has widespread appeal, and in my use of Postgres
(Postgres for websites) the need for full-text database search is
practically universal. However, it's making less and less sense
for me to support Postgres at all because so many of my licensees
can't run it with their existing hosts (who run Postgres, but
won't support a search module). Full text search in the native
distribution would almost certainly bolster Postgres advocacy.

2. MySQL supports full-text searching (not to mention
commercial-brand dbs like SQL Server 2005). Now that MySQL
supports transactions as well, there's less reason than ever for
web developers to build their applications with Postgres support.
Full-text search in Postgres (which as a contrib module is already
more functional than MySQL's integrated search) could make a huge
difference in database choice. Why be left behind when everyone
else is providing this functionality to their users?

3. Full-text search is just another, sophisticated way of querying a
database. I don't see why this can't be integrated and made to be
efficient like you would with any other common query.

4. It may an issue of efficiency vs. bulk, but really it's at least
partially an issue of efficiency vs. widespread appeal. It's true
that you want to keep the core of any application as powerful as
possible with as few unnecessary widgets as possible - but this is
something that I think Postgres needs to be viable for
consideration in larger licensed web applications.

I posted this first to the the advocacy list because first and foremost
I think that's what this is - an advocacy issue. I think that full-text
search integration is important to getting the marketspace that's being
used by what I think is Postgres's main competitor for web apps - MySQL.

Development-wise may be another issue entirely.

- Filipe

Joshua D. Drake wrote:

> Peter Eisentraut wrote:
>
>> Am Montag, 5. Dezember 2005 20:09 schrieb C. Filipe Medeiros:
>>
>>
>>> I think a major step that the development community behind Pgsql
>>> could take to advance the project would be to take Tsearch2 or OpenFTS
>>> and make one of them (I prefer the Tsearch2 architecture, but then I
>>> haven't used OpenFTS in a while) native to Postgres rather than a
>>> contrib.
>>>
>>
>>
>> This is certainly a completely misguided solution to your problem.
>> Imagine what software packages like Linux, Perl, or Apache would look
>> like today if the response to every "my hoster has not installed
>> module X" had been "let's put X in the core distribution". One could
>> certainly argue about the merits of full-text search in particular,
>> but that will only postpone your real problem until next week.
>>
>
> Not to mention that Tsearch2 is part of PostgreSQL. If your hoster
> doesn't realize the benifits, change
> hosting companies ;)... Or ask them to install postgresql-contrib.
>
> Joshua D. Drake
>
>
>
>> ---------------------------(end of broadcast)---------------------------
>> TIP 1: if posting/reading through Usenet, please send an appropriate
>> subscribe-nomail command to majordomo(at)postgresql(dot)org so that your
>> message can get through to the mailing list cleanly
>>
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings
>
>