Re: [Cert] Re: PostgreSQL Certification - Mailing list pgsql-advocacy

Hans-Juergen Schoenig a écrit :
>>
>
>> I want DBAs to be certified, not a certificate of PostgreSQL hacking.
>> Sure, PostgreSQL hackers are good at PostgreSQL DBA... But companies
>> don't want PostgreSQL hackers to manage their databases, they want DBAs.
>
> Somebody who issues a certificate has to be know more than just the
> syntax of pg_dump.

Who did say the opposite ?

> If we want to make sure that this certificate is worth anything, we have
> to make sure that people who are allowed to issue it are more than just
> "a little postgres".

We surely want certification to be a proof of a good knowledge. That
does not mean people will have to know the code in every single line.

That is what certification is suppose to verify. You certify what you
put in the certification.

But if you think we can't take the risk to give that certification to a
guy that doesn't deserve it, maybe we should think about a certification
that is something like "a core hacker will work with you for 6 months as
it is the only way to be sure you won't ever tell something that could
hurt the good feeling people have about PostgreSQL."

> otherwise we end up with a nightmare - 50000000000 people will issue
> worthless certificates.

I wondr how many centuries we'll have to wait to have 1% of that number
intersted in PostgreSQL :-)

> if every stupid guy wearing a tie is allowed to issue this, it is
> worthless ...

I'm wondering if wearing a tie and having passed SRA certification means
I'm a stupid guy?

>> I disagree completely. You'll have less than 10 companies certified
>> worldwide then. And less than a few hundred of ceritified "PostgreSQL
>> hackers", what the industry really don't care about.
>
>
> this is exactly my point.
> how many companies do you know who are really good at postgres?

Who can say who's good, by now ?

> personally i am really fed up of "self announced postgres" people.

Is there another way of proving it ?

> i am fixing problems caused by those people day after day.

We all do ne day or the other. I'm sorry, I had never coded a single
line in PostgreSQL. Does that make me a liar when I say I have some
knowledge in PostgreSQL ?

> if we decide to introduce a certificate we have to make sure that it is
> really professional.

Sure.

> to me doing some advocacy posting is just not enough to certify.
> more deep knowledge is required.

Sure.

> somebody who has been in professional postgresql business has written at
> least one piece of useful code which has been released to the public -
> so it is no big deal.

Too restricting.

> i just want to avoid that every guy who is able to type pg_dump can do
> the job ...

That is not so bad to already know how to use pg_dump, after all.

>> A PostgreSQL certification relates your knowledge in managing PostgreSQL
>> databases, not how PostgreSQL is coded.
>
>
> i am not talking about the guy who gets certified.

I thought you were.

> i am referring to the people who are allowed to issue certificates.


That is another aspect of the question.

First one is :

- What do we put in ?

Second one is :

- Who will write the certification questions. And wo will verify the
answer ?

I'm afraid that it has to be a little more automatic. And then, an MCQ
form is some good way of proceeding.

> if you know how PG works inside - then you are qualified to certify
> other people.
> if you don't have detailed knowledge about internal algorithms, you are
> just not qualified to certify other people.
>
> postgres is the database of professionals ...
>
>
>>
>>> if we don't ensure top quality, the entire thing is worthless. if every
>>> half-professional is allowed to certify, we can already stop before we
>>> start.
>>
>> I know thats not what you wanted to tell us, but Its a kind of rudeness
>> to me reading this.
>>
>> Saying that top quality is only achieved with PostgreSQL coding, and
>> that all others are half-professional is an insult to me.
>
>
> you got me wrong. i am just telling that we need strict guidelines and
> strict tests.
> otherwise it is worthless ...

We all agree. So the questions ramaining are :
- what should be in the certification program
- is an MCQ significant enough ?
- are there people willing to watch every attendee answers or is
possible to have an automatic processing of answers ?
- who will decide who's smart enough to write questions and test ?



Regards,
--
Stéphane SCHILDKNECHT
Président de PostgreSQLFr
Tél. 09 53 69 97 12
http://www.postgresqlfr.org