Given that ingres is about to spruce up an ingres janitor programme, how
many people does ingres think are waiting in the wings to become a janitor
(what an awful term)?

How many people are likely to be involved in developing for ingres in this
way?

My thoughts are that those who are likely to be involved as an ingres
janitor, are already involved in some way with ingres development/ingres
open-source, so in that way a janitor initiative will make little difference
to ingres open source development. From what I've read here, the ingres
source isn't distributed with the test tools used by ingres themselves, nor
the extensive test cases used to prove that the software is functioning as
expected for a given build. It seems to me that the effort required to
become a janitor is significant and I haven't noticed anything that leads me
to believe there's a long queue of people waiting to sign up.

Is an ingres janitor programme really going to work? Is the programme going
to cost ingres far more than the potential benefit?

For ingres, yes it's a positive news story, but I really wonder if there's a
better way to invest ingres effort.

That might wake some people up on a friday.. ;-)

Paul

Hi Paul,

These are good observations to make, and so my reply isn't a flame....

IngresCorp is well aware that the number of people waiting to do this is
probably quite small and that their skill set will initially require
them to tackle simple things under direction from a mentor.

But its not true to say that the people lining up are already Ingres
developers in some way. In my case at least, I'm just a DBA with a
curiosity about the way the code works under the hood. I've played with
OME functions and now I'd like to look at extending my knowledge and
making a contribution (however small) to the code stream. I've even (cue
stirring music) submitted my first patch! tups_per_page will hopefully
never execeed 511 again!

There definitly needs to be an improvement in the way the source is
distributed and managed when we start talking about Opensource
development. IngresCorp are well aware of this fact and have recently
made considerable efforts to improve all of this. I'd be hoping that by
the end of the year you'll see marked improvement in access to the
source code, access to cross referencing tools to find the 'bits of
code' you are interested in and some enhanced documentation to explain
it all. Check the Wiki out on a regular basis.

Furthermore, with access to mentors to assist, you'll be able to bounce
ideas off a forum and get suitable direction on how to proceed - and
even more importantly at times - why not to proceed! Personally
speaking, if I'd have had more direction a few months ago, I wouldn't
have put so much effort into trying to build a new aggregate function.
My approach was close, but not close enough. Since then I've learned why
it's a *MUCH* bigger job than I anticipated.

So yes there is considerable effort involved, it's unpaid work, and it
won't make the headlines. But there are intangible rewards. A sense of
accomplishment, kudos in the group, and an entry on the CV!

Even if you don't contribute code, why not contribute ideas for
enhancements you'd like to see? If we develop a pool of 'Janitors' you
may get much faster turn around on small enhancements than by the
traditional DAR system. And it's in relation to that last point where I
believe that IngresCorp (and you) will see the real pay-off. My
experience with DAR's was they were a blackhole. Requests went in...

For example, my next project is to look at making and implementing a cbf
entry default_index_page_size. I havent discussed this with anyone at
IngresCorp yet, but in discussion on info_ingres a while back a few
people sounded keen. And it looks like FUN!

A tracepoint to reset dm420 cache statistics...Now that's a thought.

A default index structure configuration parameter...How hard could that
be?

So what would you like to see in the code?

Martin Bowes
--
Random Duckman Quote #75:
Duckman: Alright, come-on. We don't even have a plot yet!

-----Original Message-----
From: info-ingres-bounces (AT) kettleriver...ting (DOT) com
[mailto:info-ingres-bounces (AT) kettleriverconsulting (DOT) com] On Behalf Of On
Web
Sent: 16 November 2007 09:24
To: info-ingres (AT) kettleriverconsulting (DOT) com
Subject: [Info-Ingres] ingres janitors

Given that ingres is about to spruce up an ingres janitor programme, how

many people does ingres think are waiting in the wings to become a
janitor
(what an awful term)?

How many people are likely to be involved in developing for ingres in
this
way?

My thoughts are that those who are likely to be involved as an ingres
janitor, are already involved in some way with ingres development/ingres

open-source, so in that way a janitor initiative will make little
difference
to ingres open source development. From what I've read here, the ingres
source isn't distributed with the test tools used by ingres themselves,
nor
the extensive test cases used to prove that the software is functioning
as
expected for a given build. It seems to me that the effort required to
become a janitor is significant and I haven't noticed anything that
leads me
to believe there's a long queue of people waiting to sign up.

Is an ingres janitor programme really going to work? Is the programme
going
to cost ingres far more than the potential benefit?

For ingres, yes it's a positive news story, but I really wonder if
there's a
better way to invest ingres effort.

That might wake some people up on a friday.. ;-)

Paul


_______________________________________________
Info-Ingres mailing list
Info-Ingres (AT) kettleriverconsulting (DOT) com
http://www.kettleriverconsulting.com...fo/info-ingres

Ingres Janitors is less about there being an existing queue of developers
than making sure that as people show interest they aren't turned off and
frustrated by the process. It's about lowering the barriers to entry to
getting involved. So we've talked about providing test suites, as well as
better ways to distribute source, track bugs/changes and so on.

As for effort vs benefit - I think it's going to be a slow process initially
that will gradually snowball. As it does grow it'll take less effort from
IngresCorp to maintain. But some of the things we're doing - like having a
wiki page where people can share tips and gotchas about the build process on
particular platforms - are not a huge amount of effort. But even where more
effort is needed, there is a will to do this because fundamentally we are an
open source company. We haven't always looked like one - partly due to
inertia and the history of having been proprietary, but I think that too
will change.

--
Paul Mason

"On Web" <a@b.com> wrote

First off let's talk about that term, because you are right, it looks awful.
The initialive is actually called the Ingres *Apprentices* program. The
purpose is to provide Ingres Corporation with a framework for making
assistance available to the nascent developer community. Under the program
senior Ingres management have agreed to commit a fraction of the internal
development team's time to mentoring and assisting newbies. That is
essential, because we might be volunteers but those people aren't and they
need to be paid and their time managed.

The Apprentices are expected to undertake a number of projects, only one of
which has been agreed so far, and that is the Janitors project. Being a
janitor doesn't sound very grand, and it's not meant to be. The sorts of
things janitors might do is things like systematically correcting sloppy
pointer casts. The main sense of accomplishment will probably come from
just getting a successful build.

Unfortunately, I think in our eagerness to get started with more ambitious
changes, like adding new (albeit minor) features, these are being suggested
for the Janitors too, and that's probably not right. Those will be
apprentice projects in their own right.

We're about to find out. The current planning is mainly about lowering the
entry bar. That means more and better instructions and documentation,
possibly supplying a development appliance along the lines of Icebreaker,
providing navigation aids like LXR or Doxygen, and most importantly, getting
would-be community developers hooked up with some mentors.

I suspect that's not true. The step up to get involved is currently way too
high.

Spot on. We're aware of the problem and is actively being addressed.
Continue to watch this space. Expect it to take a couple of months still.

You are so right. Please tell us more about what you think needs to be
done. The window of opportunity is wide open just now.

I don't know. I suppose it depends where you look anyway.

Most of what Ingres needs to spend money on needs to be done anyway for some
of the more professional collaborations. It's hard to see much of a
downside unless we pick up some really needy recruits. A huge part of the
reason for wanting to use c.d.i. as the communication channel is to spread
the burden helping them.

It's hard to imagine a better investment. Please feel free to make
suggestions though.

Very public spirited. Thanks. :-)

Roy

"Martin Bowes" <martin.bowes (AT) ctsu (DOT) ox.ac.uk> wrote

No, I don't read it as a flame.

Now that sounds expensive for a small payback.

Fair point. The problem I have is that while this is all good news for an
ingres enthusiast such as yourself, I'm not convinced it's to ingres
advantage.

Where is this pool of janitors coming from? Worse still, where are the
mentors coming from?

Actually, very little. I think ingres enhancements should be primarily
focussed towards allowing the DBMS to interact with other development
technologies (I have banged the drum a few times about Ruby) easily, rather
than updating ingres with finer points for ingres technologists. Similarly,
ingres should be moving to make itself the cheap alternative to it's big
brother competitors and take some of the ground between the industry giants
and MySQL.

It's good to see the enthusiasm in your post for ingres, and I see that
elsewhere in the regular posters here there are people who are enthusiastic
for the intiative.

My problem with the initiative is that it represents a cost to ingres in
terms of diverting development effort to support the initiative, which is
very likely to give a poor return in terms of ingres development payback.

I don't believe that chipping away at improving the ingres technical core is
fundemental to the success of ingres. It won't boost sales or promote ingres
in the wider world.

I would much rather that ingres investment was in initiatives that
potentially help sales and awareness of the product and I cannot see that
the janitors initiative can do that. In many ways the initiative makes me
think of what has always been a mindset problem with ingres promotion and
development all along - that technological introversion will make a product
successful.

The world demands that products interwork with each other and software such
as ingres is no longer a solution in it's own right (as it used to be) but
part of a bigger development pipeline.

Paul

On 16/11/07 12:13, On Web wrote:
the code area being worked on. As time goes by this would be expanded to
anyone who knows the code area. From my point of view, if you have
rights to submit code to a particular area, say the optimizer then you
are qualified to mentor, if you wish to that is. Even someone who used
to work on the code might be capable of providing guidance, again if
they wish to.
were aware it was available?
for Ingres Corp to implement all the functionality it needs but in the
long term we need to stir up outside interest. As has been alluded to,
the infrastructure we have now for supporting external submissions from
other companies is not up to much, the back of packet of ciggies could
be seen as an improvement. Extending the required infrastructure to
include people who want contribute (in any way) will incur little extra
cost, IMO.


--

Grant Croker - Software Engineer - http://ingres.com
MELTON CONSTABLE (n.)
A patent anti-wrinkle cream which policemen wear to keep themselves looking young.

"Roy Hann" <specially (AT) processed (DOT) almost.meat> wrote

If we imagine there is a pool of free development effort out there (and
there is) I can't see what is going to attract that pool towards ingres and
there are plenty of other interesting technologies out there for them to be
involved with.

If there was a clear trend that people are keen to board the ingres
development train, I'd support the programme, but without any evidence that
the demand exists (or can be built) then I think it will cost ingres money
and worse still divert effort away from initiatives that can make money or
build market share.

I think development effort should be driven from two directions - supporting
existing customers with bugs, etc and secondly adding innovations that will
help ingres build sales and market share.

Ingres is a seriously un-cool piece of software and it's vital that that
perception needs to change. We can all laugh that such a trivial notion as
'coolness' is important (after all it's all about keeping important data
safe, right?), but it's so important that ingres is seen as player in the
market. It's lost a lot of it's credibility against it's old competitors and
needs to establish itself as a strong solution in the market. Having lost
vast amounts of market share it needs to re-establish itself as an up and
coming technology, not something insular.

I'd like to see development effort move towards evolving technologies and
make ingres fit like a glove into those development pipelines. If you can
get the product involved with such technologies you can piggy-back off
publicity these technologies attract and re-assert ingres as a vibrant
company and solution. I think ingres needs to be in the ground that sits
between SQL Server/Oracle and MySQL. It's all very well saying that MySQL
isn't as advanced as ingres - I just wish ingres were as successful in the
market as MySQL (I just checked - turnover for ingres and MySQL is neck and
neck - circa $51Million. But their market perception is vastly different).

The future of ingres is not dependent on maintaining the status quo with
existing customers, but grabbing new customers in collaboration with other
vendors technologies. Where are the ingres and silverlight/ingres and Adobe
flex/ingres and Ruby stories?

Can anyone tell me the USP for ingres?

Paul

"Grant Croker" <grant.croker (AT) ingres (DOT) com> wrote

Hi Grant,

I haven't checked it out. If I remember correctly there were a few things
missing that kind of put me off (that and a lot of other stuff to do). I'll
look again when I've some time.

I need to check out the ingres site more often - it looks like things are
beginning to move a bit. A number of years ago I remember Business Objects
were not
particularly interested in ingres support and it took a bit of work to get
BO to work nicely with ingres, looks like that is changing.

Paul

At 12:03 PM +0000 11/16/07, On Web wrote:
I think you miss a crucial point: this *IS* an initiative that
can make money and/or build market share. The Apprentices /
Janitors project is a motivator to get the whole open source
community thing going for real.

Ingres Corp has positioned itself as an Open Source provider. From
day one, the open source-ness has been a key marketing hook, and if
you take that away or ignore it, it's very difficult for me to see
what you replace it with. And yet, Ingres Corp has done very little
to date to demonstrate any level of openness or community building
other than dumping snapshots of the source code into the public
from time to time.

We at Datallegro continue to do active Ingres development,
and while we work closely with Ingres Corp and have an
excellent relationship, the mechanics of code contribution
have never really been worked out to anyone's satisfaction.
We've managed to get some of our fixes moved back, but both
sides agree that what we are doing now isn't practical
long term. If Datallegro has trouble contributing code, with
me as an ex-Ingres'er and both sides wanting it to happen,
how is an interested outsider like Marty going to succeed?

So, to come back to your argument, I don't see this
effort as a distraction, or some nice way of patting
a few people on the head. Nor do I think that Ingres Corp
is indulging in any fantasy of raising a linux kernel-like
community of developers world-wide. *BUT* if they are
to be an Open Source company, they have to act like it,
and this is the first belated step. Trends or projections
or potential developer counts have nothing to do with it.
This, or something like it, is a MUST DO or Ingres has
to find some other "hook" besides Open Source. And the
latter is going to be a *lot* harder and more expensive.

And by the way, I fully understand the many reasons why
this didn't happen on day one, or day 365 even. It's
like trying to reconstruct a major traffic artery; there's
a zillion reasons why you can't do it today, but eventually
you have to bite the bullet. For the Ingres developers
there may be a certain amount of temporary inconvenience,
but the improvement will be permanent.



Actually, in the long run, this is another good reason to take time
now to develop the community. By the time an "evolving" technology
is evolved enough to be worth putting scarce and expensive
development resources into, it's not "evolving" any more.
Ingres doesn't have enough internal cannon fodder to throw people
at every random buzzword in hopes that it's a win. Many of these
trendy things are going to fade out. (and I sure can't predict
which ones will win and which will lose.)

I think Paul Mason said it well: it's about lowering the barriers
to entry. If we make it easy for someone who's fooling with Cool
New Technology X to pull an Ingres and plug it in, then everyone
wins. Otherwise we have to wait for X to succeed enough for
Ingres Corp to devote resources to it, and that's not going to work.

--
Karl R. Schendel, Jr. schendel (AT) kbcomputer (DOT) com
DATAllegro Inc
Ingres DBMS Server Development

Hi Paul,

Thank you for your interest. It's obvious you care or you wouldn't
have come here. Constructive criticism is especially appreciated.


I'd like to build upon what my colleagues (inside & outside of Ingres)
have said and hopefully add to the discussion you've started.


You're right, there are interesting technologies out there and we're
competing with them. We know this and don't think for one instant it's
going to be easy.

Here's our simple plan to make Ingres more attractive:


Step 1: Remove as many barriers to developing, testing, and using
Ingres as we can

This includes solving tooling issues, making our discussions more
public, providing training and mentoring, providing reasonable
opportunities for interested parties to get their feet wet and work
with us.

We can measure this. It must take less time and effort (clicks,
commands, etc.) to get an environment up and running with Ingres than
with our competition.

If we make Ingres easier to develop one (and use) than our
competitors, we've done our jobs well. This isn't going to happen over
night. We'll get there with a series of smart priority calls, constant
feedback, and hard work over time.


Step 2: Go beyond some of our competitors in terms of meritocratic
opportunity

We'd like to enable community to work along side our engineering and
support teams. Success is the day when a good submission is a good
submission regardless of where it came from. This too is measurable.

To move towards this vision, we're going to post ideas, bugs, and
other opportunities. We encourage anyone to review the list (right now
on our wiki, soon in Trac) to select something they'd like to take on.
The process is very simple:
a) Pick something you feel you can take on.
b) Review your proposal with an expert (be it an employee or community
member) to get pre-approval.
c) Once you have the approval, do the work editing/compiling/testing
d) Submit the patch for review (possibly do some updates based on
feedback)
e) The patch is committed to the code base
[rinse and repeat]

We're grateful to have bug reports and feedback on how the software's
working just as much as we're grateful to have patch submissions.


Step 3: Do work in interesting areas

This speaks to another one of your points.

If we haven't done a good job broadcasting it, we are developing
technologies with Eclipse, OSGEO, Trac, LXR, Jasparsoft, Hibernate,
and many others. In my opinion, we are doing very cool work, we're
just (possibly) doing a poor job of sharing the fact we're doing it.

Thoughts, comments, and suggestions are always appreciated.

Andrew