Re: off to the farm to start new career ...

From: Tuomas <>
Date: Wed, 15 Sep 2010 13:04:56 +0300
Message-ID: <4c909a47$0$14440$>

On 14/09/10 20:05, joel garry wrote:
> On Sep 14, 6:22 am, Tuomas<> wrote:
>> On 10/09/10 20:19, joel garry wrote:

>> Even then you have the option to learn it yourself, i.e. increase knowledge.
>> It's a real PITA, I have admit that. But it's possible. ;)
> I maintain that possibility is highly arguable. When was the last
> time you saw an application programmer delve into database internals?
> Even if they figure it out, they're likely to do more damage than
> help.

A DB is an application and it's makers application programmers or at least that's my dictionary tells me.

I'm afraid my skills in English aren't good enough if you want to argue with it.

>> Currently it also seems that it will be a drop-in replacement of MySQL, so
>> it sholdn't generate any problems.
> Uhhhhhhhhhhhhhhhhhhhhhhhh............. HAHAHAHAHAHA!
> At least you didn't say nosql. (I haven't had a chance to see the OP
> link yet).

Ahh, I see you are ignoring what's happening. And haven't thought this out.

Tell me why any capable programmer couldn't make a drop-in for a program which source code (especially file-IO) can be used as it is (recycled) and documentation about file formats is public?

Heck, even I could do that: just grab newest version from svn and off you go: you have your own fork. A drop-in replacement. And it took about 10 minutes. A couple of days to build compiling environment, though.

But that's general, this MySQL case goes even further: Do you know what happened to the MySQL-team after they (all key persons) left Oracle?

I know, they (as a team) started a new project/product. Do you want to guess what's its name?

Now, tell me a plausible reason to believe why the same guys (more or less whole team) who made MySQL in the first place couldn't do it again? Especially when they have the source on their hands and no copyright/NDA/"intellectual property rights" to stop them.

I couldn't imagine any. If you can't either, your laughing seems to raise out of ignorance.

That's one reason I like open source projects: They are impossible to kill by buying a (any) company and I'll bet that Oracle tried to do exactly that without realizing what would happen.

Totally disregarding the properties of open source software as irrelevant, especially the "open" part. And got burnt.

You could kill a project by hiring all the key persons to do something else, but can you do that? Some people can't be bought with (reasonable amounts of) money.

>>> Maybe my view is too skewed, but I see an order of magnitude greater
>>> costs.
>> From the experience or as a vision?
> Experience. I work on some closed-source stuff that supplies the
> source of the apps, and has a pool of experienced programmers and
> analysts about. Even so, after some period of time, the programmers
> and analysts go do something else. Just now I'm working on such a
> program, from 9 years ago. If it wasn't me fixing it, who would it
> be?

Someone who has the ability of learning what have done in the past. If properly documented, that shouldn't be too hard.

>The vendor still provides support, but actual maintenance work
> would be charged 10x my salary, plus additional offshoring
> communication issues. The people who did it 9 years ago? They've all
> moved on, gotten real jobs or retired, with no fresh blood coming into
> that community in this country.

And no new guys can't learn? Even complicated systems are made by other programmers and supporting them is much easier than to make new one. A big headache for the new guy, sure, but it's possible.

> With open source database or OS, the issues are more technical and
> obscure. Sure, plenty of people now, but ten years on? Even the
> current people working on current stuff, it's just unrealistic to
> think any given end-user company is going to find someone to go deep
> in a timely manner.

Unless they make a strategic decision to hire some people who know and do the support in-house. I didn't say it would be easy (nor cheap), but it's possible and it has been done.

When/if your support vanishes, you're always in trouble, as you said. The difference is "no support", "extremely expensive support" or "in-house support" and what's feasible is something which has to be decided case by case.

A small company can hire gurus who won't ever enter into a big organization, that's an advantage a small company has. An example: We've discussed with one of the guys who developed MySQL (he happens to be a Finnish guy) and it would be totally feasible to hire him.

How high level of support you want? Like using original developer(s)?

That's one of the properties open source. At least currently.

> I've seen a number of companies toss out bespoke systems that take a
> dozen programmers years to not complete, in favor of COTS. This
> usually happens when they realize they're spending a million bucks a
> year, to reinvent a $200K wheel.

Reinventing the wheel is one of those dark sides of the open source people and some wise guys invent a square wheel. Some copy stupidities from commercial programs as good ideas.

Oh well, you can't have it all.

>> Our company is now about 10 years old (most people have much longer
>> experience) and we haven't had such yet. Maybe it will come, who knows.
> As with anything, it depends. If you are in Finland, the situation
> may be better for open source than here in the US.

Hard to say, but I can say that Linux is quite popular here. ("Popular" in relative terms of course.) I'd think it has something to do with that that Torvalds is a Finnish guy and quite known, too. Almost a celebrity. ;)

One reason was XP which looked very stagnant when Fedora brought a new version every six months. New and shiny version always attract people. ;)

Even when there really isn't anything new.

> More than 10 years
> ago I was an open source advocate, now not so much.

I'd say I'm not an advocate, but propaganda against it is something I'll comment as user: There's enough to criticise without inventing it (like MS is doing).

> It appeals to me
> on an intellectual level, but like so many things, how it works in the
> real world varies.

Yes, this is something I agree with you, it's largely dependent what you do and how you do it.

Web application developing is one of the easiest tasks to do on OSS tools: Apache, PHP, MySQL. Some shell scripting and C to glue them.

Tuomas - '63 typ14, '65 typ34 & '61 typ2
Received on Wed Sep 15 2010 - 05:04:56 CDT

Original text of this message