PostgreSQL Windows port strategy

Started by Jean-Michel POUREalmost 23 years ago8 messages
#1Jean-Michel POURE
jm.poure@freesurf.fr

Dear Friends,

As a minor contributor to pgAdmin, I would like to express ideas as regards
the Windows port. As a personal point of view, it may or may not reflect the
community ideas, who knows. Don't flame me too much, I am only a casual user
of PostrgreSQL...

**************************************************************************************

Microsoft success is not due to a particular software (Word, Excel, Access, Ms
SQL Server, Internet Explorer, Visual Basic, etc...) but to the combination
of all these softwares on a proprietary platform (Windows).

Whatever free software breaks into this combination of proprietary softwares
will participate, one step after another, in the destruction of Microsoft
monopoly. Therefore, PostgreSQL and OpenOffice for Windows are probably the
most valuable efforts to break down Microsoft monopoly.

Windows alone, without Word, Excel, Power Point, Visual Basic, MS SQL Server,
will not convince a single customer. And this will be the end of Microsoft.

On the converse, Linux rise is due, not only to the availability of a large
number of softwares, but also to the notion a collaborative works inside a
community. Even without cross-porting software from Unixes to Windows, I
believe Windows will die of its own death because collaborative development
in communities is superior. It is only a question of time.

The conditions for PostgreSQL to defeat Microsoft quickly is that we stay
united and wage war on the two fronts: Windows on the one hand and
Linux/Unixes on the other hand.

Most of you are Americans and as such, you probably studied the American civil
war. One of the reasons why the civil war claimed 500.000 dies is that the
armies fought on different fronts, at different times, without
synchronization.

As regards PostgreSQL, our efforts should concentrate on both platforms at the
same time. In other words, this means: porting PostgreSQL to Windows,
bringing pgAdmin2/3 to Linux and other Free Unixies. And probably delivering
bundles under Linux, Free Unixes and Windows, offering the best Free
softwares: PostgreSQL server, pgAdmin client and PhpPgAdmin web interface.

MySQL success is largely due to its availability under Windows in bundles,
which is a pure Microsoft strategy. PostgreSQL are not and will never be
playing Microsoft strategy because we are a world community.

To sum up:

1) Microsoft sucess is due to the availibility of bundles under a proprietary
platform. Replacing Microsoft leading softwares with Free alternatives
participates in the destruction of Microsoft.

2) Linux and other free Unixes also offer a large number of softwares. Our
competive advantage is to work in communities. Even without cross-porting,
Free software is going to replace closed software, this is only a question of
time.

3) To accelerate the replacement of Windows closed source solutions,
PostgreSQL community should synchronize the releases of PostgreSQL under
Windows, Linux and Unixes. Not porting PostgreSQL to Windows is playing
Microsoft strategy.

4) Also, we should focus on offering users a bundle including: server
(PostgreSQL), client (pgAdmin), web interface (phpPgAdmin) and probably Php.
This does not need to be a single installer, but at least it should exist as
links on the web page.

Just my 2 cents.
Best regards,
Jean-Michel POURE

#2Vatamanescu Victor
victorvatamanescu@yahoo.com
In reply to: Jean-Michel POURE (#1)
Re: PostgreSQL Windows port strategy

Hi all!

Things are very unclear to me. Are we in a personal war with Microsoft and I'm not aware?

I don't really much care what's the OS our product is running on. I care much about our product's high availability, speed, scalability etc. In the last month I saw on this list a lot of opinions regarding the differences between various operating systems. I havent saw opinions regarding the functionalities exposed by our direct competitors: other dmbs. Do we want to transform PostgreSQL in the next generation's OS and I am not aware?

I can guarantee you good software can be written on Windows too. I can guarantee you that moving PostgreSQL on Windows is the best move PostgreSQL. I can guarantee you a MMC Snap-In for PostgreSQL and a Visual Studio .NET add-in for PostgreSQL can help PostgreSQL becomming a leader.

But discussions like "Windows is a B category platform" and "Windows will die tommorow" and "Linux is the best" won't lead us on a top position.

---------------------------------
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day

#3Daniel Kalchev
daniel@digsys.bg
In reply to: Vatamanescu Victor (#2)
Re: PostgreSQL Windows port strategy

Vatamanescu Victor said:

I don't really much care what's the OS our product is running on. I care muc

h about our product's high availability, speed, scalability etc. In the la
st month I saw on this list a lot of opinions regarding the differences be
tween various operating systems. I havent saw opinions regarding the funct
ionalities exposed by our direct competitors: other dmbs. Do we want to tr
ansform PostgreSQL in the next generation's OS and I am not aware?

PostgreSQL is what it is, which certainly is not an OS.

If PostgreSQL was 'Operating System' we wouldn't care much on what 'Operating
System' it runs, right?

But most of the things related to performance AND functionality of PostgreSQL
depend very much on what OS you use to build and run it on. (otherwise it
could well contain portions of the OS much like Oracle does :)

While I agree, that (any) Windows platform may be useful for a lot of things,
it's true that one should be wary of Windows, for things like software
development (due to it's unstable API), platform stability (I have yet to know
someone who didn't have to reinstall Windows every month or so), performance
(you can hardly explain to customers, why their new personal Windows
Workstation requires > 1 GHz processor, > 256 MB RAM etc while showing them
how a moderate PostgreSQL database server that could serve their entire
company can run on much older/slower/cheaper hardware).

There is nothing wrong in having nice Windows based GUI for accessing
databases, but turning a graphical workstation into database server is
something that has never worked in the computer history (the opposite has
happened from time to time).

Windows has lost the game when Microsoft decided to abandon support for
non-Intel CPUs in Windows NT (VMS actually).

I can guarantee you good software can be written on Windows too. I can guara

ntee you that moving PostgreSQL on Windows is the best move PostgreSQL. I
can guarantee you a MMC Snap-In for PostgreSQL and a Visual Studio .NET ad
d-in for PostgreSQL can help PostgreSQL becomming a leader.

While enough was said in this list regarding the Windows port - nothing could
stop anyone to port an open-source database to whatever platform, including
Windows. The better the platform knowledge of the porting team, the better the
result. But 'moving' PostgreSQL to Windows won't happen.

But discussions like "Windows is a B category platform" and "Windows will di

e tommorow" and "Linux is the best" won't lead us on a top position.

My personal favorite is BSD/OS and for what it matters, Windows cannot solve the kind of tasks, that BSD/OS solves for me. period.

Daniel

PS: I probably sound too anti-Microsoft biased, which is not true. I just don't care about Microsoft - I use Windows from time to time when someone sends me $@^#$&@*&#&*$-Microsoft-Office formatted document that will not open anywhere else. Windows also makes great launcher for my father's Heroes III game. ;-)

#4Vatamanescu Victor
victorvatamanescu@yahoo.com
In reply to: Daniel Kalchev (#3)
Re: PostgreSQL Windows port strategy

Well, I havent seen much that unstable API. If you saw something unstable please provide me source code that proves Windows API is unstable. Don't tell me about some "expert"'s oppinion: if you have a problem with Windows show it to me. We are not users here. I agree that Windows 95/98 was a sh..., but I have news for you: Windows is becomming stronger every day. In 1998-1999 we could talk about the dying Microsoft - they were null on the server platform and the growth of Linux seemed unlimited. Can you tell me that the situation is the same? Have you carrefully tested Windows 2003?

About reinstalling: at home I have a Pentium III 1 ghz workstation. OS: Windows XP. I have installed on it Visual Studio .NET, Visual Studio 6, Microsoft SQL Server, IBM DB2, Cygwin, IIS, MSMQ etc etc(you got the picture). I installed it about 9 months ago (when I bought the hdd) and I dont have any trouble with it, I hadn't to reinstall it after a month...

By the way, DB2 has no problem working on Windows. I havent heard IBM calling Windows a "B category platform", on the contrary...

---------------------------------
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day

#5Daniel Kalchev
daniel@digsys.bg
In reply to: Vatamanescu Victor (#4)
Re: PostgreSQL Windows port strategy

Vatamanescu Victor said:

Well, I havent seen much that unstable API. If you saw something unstable pl

ease provide me source code that proves Windows API is unstable. Don't tel
l me about some "expert"'s oppinion: if you have a problem with Windows sh
ow it to me. We are not users here.

Really? We ARE users of Windows here. Nobody has ever bothered to tell US how
this thing works, why it works and what can we, users do when it doesn't work.

I am talking about the Windows API here, with it's numerous variants and
'Service Packs'. I have learned my lesson on Windows: use only default
configurations, anything else will bite you badly some day.

(this is not to say that you can't get the same effect on other platforms,
just that on most 'commercial' UNIX platforms some things are very detailed
documented - sometimes too much in detail)

I agree that Windows 95/98 was a sh...

, but I have news for you: Windows is becomming stronger every day. In 199
8-1999 we could talk about the dying Microsoft - they were null on the ser
ver platform and the growth of Linux seemed unlimited. Can you tell me tha
t the situation is the same? Have you carrefully tested Windows 2003?

You are wrong about Microsoft here. They were null on the server platform, say
at 1985, when their wonderful product "Microsoft BASIC" was my favorite on the
Apple ][. Then IBM gave them the IBM PC to write an 'OS' for it.

I don't claim to be an expert on Windows, although I have actually 'tested'
(both from the user's perspective and usability as 'platform') every Windows
release since 1.03.

About reinstalling: at home I have a Pentium III 1 ghz workstation. OS: Wind

ows XP. I have installed on it Visual Studio .NET, Visual Studio 6, Micros
oft SQL Server, IBM DB2, Cygwin, IIS, MSMQ etc etc(you got the picture). I
installed it about 9 months ago (when I bought the hdd) and I dont have a
ny trouble with it, I hadn't to reinstall it after a month...

But rebooted it how many times?

I have never ever reinstalled a UNIX system. Some run for years (yes, that is
some N x 365 days) without being reboot or powered down. They usually stop
when that same flaky PC hardware breaks, and I have to build new system then
anyway.

By the way, DB2 has no problem working on Windows. I havent heard IBM callin

g Windows a "B category platform", on the contrary...

I was trying to avoid this part of the discussion... mostly because I don't
believe PostgreSQL is yet able to compete with the 'big' database systems.
Here I mean BIG database systems. One of the reasons in my opinion is that
those guys that run their data on big iron just can't explain why they will
spend millions of dollars a year on OS licenses and still want to install an
open-source database there.

It seems you got my opinion wrong: I don't see anything wrong with running
PostgreSQL on Windows. In fact, I have been considering such port some time
ago (or alternatively a 'DOS' port :), because this allows greater flexibility
in certain cases. However, moving main development effort on Windows, just to
satisfy the Windows lovers' ego would be catastrophic and simply will not
happen to PostgreSQL. One of the reasons this will not happen is because
PostgreSQL has already grown, has become quality software and many people
already know how it runs on the SAME Intel hardware on which Windows runs.

The difference with DB2 that runs on IBM other platforms is significant for
the same reason. When it fails at your data set on the Windows platform, IBM
can always tell you "but,... this is because of the cheap and flaky PC
hardware - here is what this wonderful software is designed to run on.."

#6Merlin Moncure
merlin.moncure@rcsonline.com
In reply to: Daniel Kalchev (#5)
Re: PostgreSQL Windows port strategy

I think there should be a special mailing list set up called
pg-sql-win32-advocacy where people can continually harass the postgres
dev team and debate the merits of the win32 operating system. In
particular, make sure to insult them for being elitists (while at the
same time asking them to work on software which you have not paid for).

We can also rout all the questions about how to set up cygin and the
ipc-daemon over there too.

Merlin

#7Jean-Michel POURE
jm.poure@freesurf.fr
In reply to: Merlin Moncure (#6)
Re: PostgreSQL Windows port strategy

Le Mercredi 12 Février 2003 15:49, Merlin Moncure a écrit :

I think there should be a special mailing list set up called
pg-sql-win32-advocacy where people can continually harass the postgres
dev team and debate the merits of the win32 operating system.

I realize my views about PostgreSQL are not shared by anyone, otherwize in a
few hours time, I would have received several emails saying "I have the same
feeling, etc...". It seems like everyone is looking for something different,
which ultimately turns out to become ... PostgreSQL.

#8Merlin Moncure
merlin.moncure@rcsonline.com
In reply to: Jean-Michel POURE (#7)
Re: PostgreSQL Windows port strategy

Well said
I was just trying to be funny. Since we are all programmers here, I'll try and re-express my thought in a (somewhat) portable language:

char HackerThought[] = "Maybe there really is something to this windows business. Perhaps its time to consider a port".
int iAnnoyanceLevel = 0;

PerformDailyRoutine();

PerformDailyRoutine ()
{
char* EmailMessage;
Think(HackerThought);
while (EmailMessage = GetEmailMessage(HACKER_LIST))
{
if (!strcmp(EmailMessage, "Windows is an unstable, crappy OS")) iAnnoyanceLevel++;
if (!strcmp(EmailMessage, "Windows is better than that pile of trash, linux")) iAnnoyanceLevel += 2;
if (!strcmp(EmailMessage, "Can anybody here tell me how to install the IPC-Daemon as a service on my win98 machine?")) iAnnoyanceLevel += 100;
}
}

Really, I'm new here, and I have not business telling anybody anything, anyways :)
Merlin

Show quoted text

-----Original Message-----
From: Jean-Michel POURE [mailto:jm.poure@freesurf.fr]
Sent: Wednesday, February 12, 2003 11:26 AM
To: pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] PostgreSQL Windows port strategy

Le Mercredi 12 Février 2003 15:49, Merlin Moncure a écrit :

I think there should be a special mailing list set up called
pg-sql-win32-advocacy where people can continually harass the postgres
dev team and debate the merits of the win32 operating system.

I realize my views about PostgreSQL are not shared by anyone, otherwize in
a
few hours time, I would have received several emails saying "I have the
same
feeling, etc...". It seems like everyone is looking for something
different,
which ultimately turns out to become ... PostgreSQL.

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)