Linked by Paul N Richter on Tue 3rd May 2005 21:31 UTC
Databases The mid-market is hot. It is forecasted to be the main growth segment in the IT market for the next few years. As such, it is fiercely contested by the major software makers. The BBC recently even called it a 60 billion dollar software battle. What exactly constitutes the mid-market is a bit vague, but some define it as businesses ranging from 50 to 1,000 employees.
Order by: Score:
v PostgreSQL
by Martin on Tue 3rd May 2005 21:42 UTC
nice article
by Anonymous on Tue 3rd May 2005 21:52 UTC

don't trust such marketing article.....

ERP not DB
by Jon Smirl on Tue 3rd May 2005 22:06 UTC

The article is not really about databases, it's about ERP packages.

Clutching at Straws
by David on Tue 3rd May 2005 22:06 UTC

These software companies are clutching at straws. The notion that a small to medium sized business is going to spend a spectacular amount on software, and in waste, is just a total fallacy. They don't go through the churn a large enterprise does, and these companies and analysts want to get themselves another job if they think that these companies will spend.

Oracle is totally and utterly, grossly, overpriced for such a target market and SAP - oh my God. Many people who run Oracle databases today simply don't need it, and Oracle would be well advised to wise up to that threat in the next few years. They (as well as SAP) also produce some of the worst software I have ever seen in my life.

Not only is the software of an extremely poor quality, but you're then roped into spending an obscene amount of money on Oracle, and especially, SAP consultants. It's not going to fly here and these people are living in extreme dreamland.

marketing piece for firebird...
by Anonymous on Tue 3rd May 2005 22:21 UTC

This is quite the marketing piece for firebird. I hate to say it, but when you leave postgres out of the top five contenders, and stick in firebird, that's a problem. Postgres has the edge when it comes to stability in the open source arena, and it's amazingly featureful. Especially with postgres 8 being released for windows, it should have been considered in the article as the top open source alternative to the likes of SQL Server.

Firebird?
by Matt on Tue 3rd May 2005 22:27 UTC

Geeze, I wonder if the author is sleeping with a Firebird developer. Does anyone actually use firebird? I evaluated it along with MySQL and Postgresql. PostgreSQL for our needs was better hands down. The quote from the GoldMine guy of

"Firebird came out on top based on a combination of factors including licensing, install process, performance, administration, feature set, and cross-platform capabilities"

Postgresql has all the major features of Firebird. The both have minor features that the other ones doesn't. PostgreSQL is licensed under BSD. It also complies on a ton of platforms. There pay for support available. What else do you need?

I don't know, but this "article" sounds a lot like a advertisement.

Good article.
by re on Tue 3rd May 2005 22:30 UTC

I whole heartedly dissagree with the above comments. This article is right on the mark. I have been investigating this market for some time myself and have found the exact same results.

Its also not a puff piece for any of the aforementioned products its simply a statement that thats really all that is available.

In fact the mid size business market presents some complicated issues. Its true they dont have the large budgets that the very lage businesses do and yet the choices on the lower end while getting better arent really all defined.

@Matt
by Robert Escue on Tue 3rd May 2005 22:40 UTC

Exactly, is this a plug for firebird or what? What about Oracle or DB/2, or for that matter SQL Server (Sybase)? I expected to read a comparison article, and as such it is a dismal failure.

RE: Good article.
by Matt on Tue 3rd May 2005 22:47 UTC

Sorry, but I actually read the article. It did not go into any details of WHY it is better then the other free alternatives. If gives a few examples of where it is being used, and a few people who are probably supervisors like mine who have don't know a trigger from a view.

The only thing I got from it was that Firebird has an "oracle-mode emulation layer" which I might need to use at some point.

Other then that, it is a fluff piece. It keeps saying things like Firebird has a "business friendly license", well as does Postgresql. There isn't really a comparison given between any of them. Has the author even used or read about any of the others?

Joke
by youknowmewell on Tue 3rd May 2005 22:49 UTC

This article was a laugh, but that's about all. The author of it should work on subtly a bit more and try again.

You will only see it from the point of view of a product review. And having gone throug a very similiar situation as this author I see they are right on the mark.

Although my preference is to use Mysql and then use oracle's migration utility.

Old info?
by Uno Engborg on Tue 3rd May 2005 23:42 UTC

The auther seam to prefer Firebird, and no doubt, this is a very good database. However he doesnt tell us why.

He have issues with the license. Postgresql is available under the BSD license. This makes it possible to use it in any way you like. for free or closed source software.

The crossplatform abilities of Postgresql is not a problem anymore as it now runs natively on windows.

The article also fails to mention the main problem with Firebird. The lack of documentation. Here postgresql is much better

Unless I needed an embedded database I would go for Postgresql any time.

humbug
by Ceaser on Wed 4th May 2005 02:45 UTC

if the article was about postgres instead of firebird none of you would be whinging that its "a marketing piece" etc.

its nice to see a DB oriented article NOT about mysql. at leatst it was nut trumpetning mysql-5s uber stored procs, or mysqls uber triggers.

if it _is_ marketing, its good to see firebird marketing team getting it out there in the press. too long they have not really made much noise about an excellent product.

Legacy
by Someone on Wed 4th May 2005 02:49 UTC

I haven't tried Firebird. However most database preferences seem to come down to legacy.

Postgres has a bad legacy back from version 6 amongst the MySQL developers I know of. I know its totally undeserved given the recent versions, but back then I chose MySQL and more often than not, I still do.

One of these customers has been running Firebird or similar for seven years. Postgres was crap seven years ago, even compared to other products as they were around seven years ago.

That said I would say Postgres is probably the most feature full database on the market, even more so than most commercial offerings.

I'm still using MySQL because most of the applications I work with are still using MySQL and there hasn't really been a need to upgrade. I've worked with Postgres though and found it to be more than capable, although my experiences have been plagued by what I felt were unreasonable bugs for a stable version.

In the future I'm open to all options, even MSSQL or Oracle although open source databases more than fill my requirements these days.

I was willing to put $50 that first comment would be about Postgres. It always is. (Even in articles where it was hinted that the situation was windows based.) I could be rich right now.

@Ceasar
by youknowmewell on Wed 4th May 2005 03:14 UTC

The problem I think we've had with this marketing piece is that it tried to (very poorly mind you) disguise itself as some real article, rather than advocacy or marketing material. The title was misleading as well.

Title: Selecting a mid-market database.
Content: Firebird is good choice.

You see why this material doesn't sit well with us?

the article itself was very valuable for once again bringing Firebird to my attention. While I am partial to PostgreSQL (especially now that it is cross-platform), I have seen a lot of positive reviews of Firebird.

I would say that PostgreSQL has far more language support and third party utilities available for it than Firebird, as well as better documentation and project support, but the reasons for choosing Firebird for a specific project enumerated in the article are entirely reasonable. Footprint, stability, SQL standard coverage, etc. are all valid reasons for choosing a database depending on your requirements.

Oracle and DB2 and SQL Server are useless for the small business market (which overlaps the "mid-range" market if you're talking about 50 employees) because of the enormous management issues involved which is more important than the huge license fees. Firebird, PostgreSQL, and MySQL all have their place for this market, more so than the big contenders. The higher you get in the "mid-range", the more reasonable it is to use the big contenders - IF your project REQUIRES that kind of overkill - which many projects on the departmental or divisional level probably don't.

Any time open source alternatives get positive press over the big boys, it's a net win for everybody, so chill out.

This is just the software industry synthesizing a new market for itself where one could hypothetically exist but in fact doesn't. Why would this industry bother investing in such pipe-dreams? The large-firms are sewn up, the tiny companies probably pirate software...so of course everyone in the middle is just waiting to spend $$$$ on slimmed down diluted versions of the same code the big boys already run.

WRONG. In reality the entire business software industry is in terrible shape. You have the whales (MS,Oracle), the service+software players (IBM), and the niche players (Adobe). Everyone else (Seibel, Sybase etc) are getting swallowed or stomped. There is no angel market that will flush them with cash again, because the whales are just as effective in marketing to these smaller firms and have more credible futures. And even for the whales things could be better. Oracle has had to acquire to grow. MS is not growing at all, they are just trying to hold on to what they have.

Yes there are people spending money on code in mid-market firms. No they will not all go to MySQL (many will though)...but nothing can stop the inevitable consolidation that is waiting to take out 75% of business software firms still alive today.

The article seriously considers MySQL and Firebird but fails to even mention PostgreSQL.

This doesn't make sense given that PostgreSQL has a substantially larger userbase than Firebird, and a much stronger feature set than MySQL.

That is, until you Google the author's name and realize he wrote about Firebird in past articles. Funny, those past articles mentioned PostgreSQL before it had a native Windows version available.

I suspect touting the advantages of Firebird over PostgreSQL 8 is now considerably more difficult for the author.

Those Firebird conspirators!
by Zambizzi on Wed 4th May 2005 04:49 UTC

Please, there's no clandestine effort by "the man" to keep Postgres down.

Firebird is superior in many ways...

1. Easily embedded into any kind of Windows application (single .dll library)

2. *Very* lightweight, the entire install is under 3mb.

3. Very mature, has existed for something like 20+ years, was formerly Borland Interbase, so it has had a presence in the enterprise DB market for decades.

4. *extremely* fast query engine...in my own personal and professional use, it blows the doors off of any other database that I've used.

5. Powerful and simple SQL language extentions in DSQL, PSQL, etc.

6. Simple to manage. Simple to install...postgres's install on linux is like pulling teeth, IMO.

I could go on but why bother. It's great to see Firebird getting the attention it deserves as a fantastic embedded/enterprise database. Goldmine is a terrible product, IMHO, but the market doesn't agree with me...so more power to em'!

Those Firebird conspirators!
by Zambizzi on Wed 4th May 2005 04:50 UTC

Oh yeah...

7. A *long* history on the Windows platform, Postgres just started dabbling in that area...and maturity of use on the platform could have been one of the considerations.

v xx
by xgxg on Wed 4th May 2005 08:40 UTC
In other news...
by Martijn Tonies on Wed 4th May 2005 09:02 UTC

"Does anyone actually use Firebird?"

What makes you think people don't? It's a very decent system that works well on many platforms. Reliable, easy to install and very capable.

"Postgresql has all the major features of Firebird"

Except, that until very recently, it was not at all easy to install on Windows, let alone being able to find a Windows package on its website :-)

" It keeps saying things like Firebird has a "business friendly license", well as does Postgresql."

Actually, nothing negative is said about PGSQLs license. The paragraphs where licensing is mentioned, MySQL is mentioned as well. And as we all know, MySQL doesn't come for free (anymore) for commercial applications. Firebird does. So does PostgreSQL.

"Has the author even used or read about any of the others?"

Why should he? The author has written an article about the future of mid-market-related database systems and gives a few example of companies/systems that use Firebird and let's them talk about why. A blow by blow comparison is not what he wanted or, apparently, intended.

It seems that the PostgreSQL people in the comments are a bit annoyed with the lack of PGSQL in this article... Well, Firebird has suffered from this for years - it's about time that it got some coverage in the news...


User base
by Bobby J on Wed 4th May 2005 09:30 UTC

"Does anyone actually use Firebird?"

See here:
http://www.fyracle.org/showstory.php?id=125

Firebird: 3 million
PostgreSQL: 100,000

oracle compatibility?
by andre on Wed 4th May 2005 10:02 UTC

is this something that Firebird has out of the box? its stored proc language is COMPLETELY different from PL/SQL...or is this article a *plug* as well for that janus software thing?

:P

Re: User base
by Anonymous on Wed 4th May 2005 10:07 UTC

3 million includes Interbase (20+ years developed history)...

so should we count Ingres as postgresql user base (also 20+ years developed history?)

I agree Firebird is greate DB, but lack of document.

Re: Re: User base
by Bobby J on Wed 4th May 2005 10:31 UTC

Most Interbase users have switched to Firebird. This is easy because API's are identical and dev tools work with both. The porting effort is *nil*

Have most Ingres users switched to PostgreSQL? I think not. If anything, they seem to be switching the other way:
http://www.it-director.com/article.php?articleid=12637&SESSID=1d4d7...
Also, CA says that the Ingres installed base is less than 10,000:
http://www.computerweekly.com/Article134656.htm

For Firebird documentation, see for instance:
http://www.amazon.com/exec/obidos/tg/detail/-/1590592794/qid%3D...

Or this list:
http://www.ibphoenix.com/main.nfs?a=ibphoenix&s=1115201907:509135&p...

RE: Re: Re: User base
by Martijn Tonies on Wed 4th May 2005 11:09 UTC

"Most Interbase users have switched to Firebird."

That's a bold statement to make.

Although I'm a Firebird supporter, I'd rather say that "some" or "many" InterBase users have switched to Firebird. There are also many first-time users of Firebird.

For newer versions of InterBase, depending on what you're using, the porting efforts are not NIL.

InterBase and Firebird are no longer the same.

They are very similar, but not the same. Not at the surface (SQL features, for example) and not under the surface (eg: query optimizer differences).

RE:oracle compatibility?
by Uno Engborg on Wed 4th May 2005 11:35 UTC


is this something that Firebird has out of the box? its stored proc language is COMPLETELY different from PL/SQL...or is this article a *plug* as well for that janus software thing?



It's a plug for Fyracle development kit, that is an add on to Firebird to make it reasonably compatible with Oracle. E.g. PL/SQL support.

@Zambizzi
by Steven Haryanto on Wed 4th May 2005 13:12 UTC

Simple to install...postgres's install on linux is like pulling teeth, IMO.

You are probably comparing installing FB's Linux binary vs building PG on Linux from source? Most distros already provide binaries for PG, it's no more difficult than a simple "rpm -i" command or a couple of clicks on Synaptic.

In fact, last time I tried building FB 1.x from source it's far more difficult than PG. Nonstandard build procedure, bootstrapping, etc. AFAIK, FB has only embraced autoconf in the recent FB2 codebase.

@Bobby J
by Steven Haryanto on Wed 4th May 2005 13:24 UTC

Fits with surge in popularity polls

... except that FB users (e.g. in the fb-general mailing list) are always encouraged to skew the polls by voting like crazy :-) While PG users seem to can't care less...

Despite website polls, I am still confident PG is the number one open-source database. Well, number two perhaps, if we consider MySQL. FB must be the quite distant third. Why? I believe mailing list traffic and the pace of development is a more objective indicator than website polls, don't you agree? :-)

I have nothing against FB, but when will FB 2.0Final be released, by the way? It's still experimental/alpha since forever. Since the days of PG 7.1 or 7.2 I think. We have now already seen PG 7.3, 7.4, 8.0, and soon enough 8.1 with 2-phase commit and integrated autovacuum. Those are perhaps the two biggies that FB currently has and PG lacks. We will soon see that there is practically nothing FB has that PG doesn't have except embed-ability. Sad, really. For some reason FB development doesn't take off as fast as we all want/wish it to be.

Popularity polls
by youknowmewell on Wed 4th May 2005 13:55 UTC

Judging by the responses to any DB article on both Slashdot and OSnews (large amount of people talking about PGSQL) I'd be willing to bet popularity for PGSQL is higher. You certainly don't hear much at all about Firebird unless there is some obvious marketing material like this article that talks about it.

This is a joke right
by Anonymous on Wed 4th May 2005 15:37 UTC

This is not an article, not a comparision, not even a useful description of the blatently pitched database engine. Its a pragraph of useless blabber to at least name other databases, followed by a few opinions of users with no useful detail. Why exactly was this worth posting?

SQL Server?
by Anonymous on Wed 4th May 2005 15:44 UTC

Anyone thought of SQL Server 2000? Does all of that and costs not much.

geographic user base
by Josh on Wed 4th May 2005 17:32 UTC

Postgres seems to be popular in the US and Firebird in Europe. Just look at the posts in this talkback. During US office hours people think PG is the only show in town. During EU office hours the Firebird proponents make their case.

What about SQL Lite
by Scott Marlowe on Wed 4th May 2005 18:23 UTC

Given that we're talking a contact manager which isn't likely to have a whole lot of concurrent updates, why not look at SQLLite?

While I tend to think of PostgreSQL as a good "heavy lifting vehicle" It's needs for am observant and intelligent DBA to handle configuration and ongoing maintenance may well be what made it lose out in this competition.

While SQL lite uses full table locking for updates, thus making it a poor enterprise class choice, it's ability to be embedded and run quite quickly when doing mostly reads make it a particularly good choice for content management, which is mostly what a contact manager tends to be.

Anything would be an improvement over the old gold mine, which was prone to corrupting all it's data at a moments notice back in the day.

...
by Anonymous on Wed 4th May 2005 18:41 UTC

I've tried both FB and PG, while PG is more secure and stable FB is easier to use, deploy and administrate.

When PG for Windows came out I really wanted to use it but after making some comparations between both (making test programs) I couln't choose PG, it is to slow.

I liked more the transactions structure of FB than PG.

Who knows, maybe in future versions PG will be faster and easier to deploythan FB.

Firebird
by Will on Wed 4th May 2005 18:57 UTC

WRT to Firebird bundled with Goldmine, it makes complete sense. Firebirds Interbase roots make is a solid choice to run embedded on Windows.

Firebird is also very well suited for embedding in general.

Postgres has a VERY short history running on Windows.

I would like to find a better comparison of Firebird vs something like MSDE, though, since MSDE is effectively a free database to distribute for MS developers, and its heritage is from SQL Server (for good or ill).

Actually, a Firebird vs MSDE vs SQL Anywhere would be an interesting comparison I think.

But I give Postgres more traction as an "enterprise" database for use on a central server vs the embedded market in a product like Goldmine.

Firebird Advantages over PostgreSQL
by mariuz on Wed 4th May 2005 19:22 UTC

Open source databases should target db2,mssql or oracle
That is about in article
If you really want an pg vs fb style article here is one
http://www.dotnetfirebird.org/blog/2005/03/firebird-advantages-over...
There are other comparisions on the same page
Firebird versus Microsoft SQL 2005 Express
http://www.dotnetfirebird.org/blog/2005/02/embedded-firebird-and-mi...
Firebird versus MSDE 2000
http://www.dotnetfirebird.org/blog/2005/01/embedded-firebird-and-ms...
Firebird versus Microsoft Jet http://www.dotnetfirebird.org/blog/2005/01/firebird-and-microsoft-j...

Over 1 million users? Not hardly
by Josh Berkus on Wed 4th May 2005 21:34 UTC

Paul,

While I appreciate Firebird getting good press (and it certainly deserves more than it has gotten, historically), there is absolutely no way that it has an installed base of more than 1 million. Perhaps if you include everyone who ever was an Interbase user ever, you could reach 1 million but that's the only way.

This article would have been fine if it had been pitched as "Firebird Database Popular with mid-market Application developers", but as a comparison of mid-market databases it leaves a lot to be desired; there's no comparison, and only one database is discussed. Maybe ask someone to give it a read-over next time?

--Josh Berkus
PostgreSQL Project

RE
by Guillaume on Wed 4th May 2005 22:33 UTC

The thing with firebird is delphi...don't forget that a lot of people still using that to develop application, and the support with firebird is easy.
Same with .Net / firebird.
I have really no clue if PostgreSQL can be used that easily to develop relatively fast different applications ( I hope that I make sense >< )

Josh,

Naturally, when it comes to statistics/counts, they are like benchmarks -- everybody can make any number make sense.

In general, however, the project calculated the 1 million number based on:

- Statistics from the SourceForge project site, which shows that almost 2 million downloads have taken place. (BTW, the SF site stopped updating the stats in Jan 2005 due to a SF issue)

- There are a number of local mirrors for the Firebird kits, which are not reflected in the SF counts and for which we don't get statistics.

- there are also CD which are available which have the install kits as well (but this is not significant)

You will note that the Interbase isntalls are not included.

So, by factoring down the downloads by 50% you get 1 million real downloads/users. Again, we can argue about the numbers but that's not going to accomplish everything.

From there we can get into the debate of "what is a user?".

Personally, I think it is people who use systems which use the engine -- in our case, we have over 300 users of our software -- from the single download.

And the 1 million installs, that's very simple. Although, in my case my firm has over 25 sites running our software, there is a estimating factor which is used by industry consultants like IDC which provides 10 installs per known download of Open Source software.

Finally, I also agree that the title of the article is a little misleading.

Sean Leyne
Firebird Project

What about 64 bit support
by Richard on Thu 5th May 2005 05:44 UTC

While the embedded use of Firebird sounded interesting, problems with 64 installs forced be to install PostgreSQL on my Gentoo-AMD64 laptop. Given databases are one of the application that benefit for 64 bit processors I would have thought fixing this would be a high priority fix.

Thanks + Installed base
by Paul Richter on Thu 5th May 2005 08:09 UTC

First of all, a big thank you to all who provided feedback. I had expected most feedback to be from SQL Server afficionados, not from Postgres afficionados. After all Postgres is mentioned 3 times, and never in a negative context. I am still surprised at the absense of comments from Oracle, DB2, SQLServer, MySQL, Sybase and Ingres users.

Postgres posters seem to have two complaints, the title and the installed base. I think the title is valid if you look at it from the viewpoint of a midmarket developer, but I can see how people can find it misleading. My appologies for this.

Using very different reasoning than Sean (see above), I think that the 3 million claim for Firebird+Interbase could be correct. As it is so controversial, I am planning a follow-up piece on the estimated installed base of production installations of Oracle, DB2, SQLServer/MSDE, MySQL, Firebird/Interbase, PostgreSQL, Ingres and Sybase.

I asked about the number for PostgreSQL on the pg-advocacy mailing list, but never received a reply other than that the moderator was reviewing my post. So far, I think the 100,000 number might be in the right ball park as well. My purpose is not to talk down Postgres, so Josh, I would appreciate all help I can get from the PG community with getting numbers for any and all of the systems listed above.

That same question goes out to the MySQL community too. The MySQL CEO is on record as saying that MySQL has an installed base of 5 million in 2005 and 6 million in 2006. For that statement to be meaningful he must know the number to within 20% accuracy. Perhaps MySQL has "phone home" code built in to measure actual use, perhaps they have another very accurate estimation method; I don't know. My current estimate for MySQL would be 1 to 3 million.

Paul

RE:Firebird Advantages over PostgreSQL
by Uno Engborg on Thu 5th May 2005 10:58 UTC


If you really want an pg vs fb style article here is one
http://www.dotnetfirebird.org/blog/2005/03/firebird-advantages-over.....


It seams that postgresql is closing in. In a couple of years Firebirds oldest in windows town advantage will fade, and soon nobody will remember windows 98.

The licensing issues with the LGPLed .Net stuff in Postgresql would affect very few businesses as most companies don't intend to develop databases , they intend to use them, and as long as they do that, they are free to make closed source solutions.

Add to this, that postgresql development is much faster. I would expect that we will see postgresql 9.0 with lots and
lots of goodies before a we se a stable Firebird 2.0 release.

The postgres firebird numbers are weird
by Werner on Thu 5th May 2005 12:36 UTC

First of all, while I really like firebird, the communities around the project and the projects history show that the numbers are weird.
Firebird has a rather small community (while the DB is very solid indeed but lacks a macosx build and decent non windows admin tools), while the PostgreSQL community is growing rapidly.

Secondly, the windows build from PostgreSQL was downloaded roughly around 80.000 times alone, indicating that the user base of postgres has grown due to the windows build as well.
Third, Firebird counts in with teh 5 million userbase the people who basically got interbase via delphi, jbuilder and other Borland products, but yet fail to count in that Postges comes as defacto install with every bigger Linux distro there is. Redhat alone might count for several millons of so called users (which mostly are not even aware that there is a full blown enterprise SQL server on their box, like most Delphi users arent aware of Interbase, thanks to the Borland docs which dont put much emphasis on that fact)

All in one all this pickering is pointless, Firebird is an excellent DB and if I would have a Windows application which needs embedded DB stuff I would hop onto Firebird in an instant, for enterprise stuff, I nowadays would go to PostgreSQL, thanks to the good cross platform tools and its nice license and the huge momentum it currently has.
MySQL is dead for me, it never really was a viable option and sort of a joke of history, that yet again the worst solution there was, and there still is has the biggest market share. (That will hopefully will change soon)

Blobs
by N.N. on Thu 5th May 2005 15:39 UTC

is directly streaming stored blob pictures to the customer browser.

This is overkill. There is not practical benefit of storing images in the db.

postgresql users
by Scott Marlowe on Thu 5th May 2005 16:23 UTC

Paul Richter, I doubt there are only 100,000 postgresql users world wide. After all, apple uses it underneath their desktop nowadays if I understand correctly. There were 100,000 downloads of the Windows version of pgsql 8.0 since it was introduced. It was used as the standard control database underneath the RAQ / Qube systems by cobalt. I'm sure it's stuck inside all kinds of other things too.

I imagine the number being firmly into the millions, but more importantly, who cares?

PostgreSQL, like good science, doesn't rely on popularity. it relies on quality. Review how well each works, for different workloads, and make it clear that the number of users is basically something NO ONE can really count, as it's quite possible to get it from many sources, and no one's counting anyway.

More follow up on numbers
by Josh Berkus on Thu 5th May 2005 16:43 UTC

Sean,

Thanks for explaining the numbers. If the PostgreSQL project uses the same calculation you do, then we'd get an installed base of 2 million ( 200,000 downloads of version 8.0 x 10 ). Which I'm happy to promote if I can cite you as a source. Deal?

BTW, I never believed the MySQL numbers either until the MySQL Worm hit, and over 150,000 servers were down. Marten won that argument on that day.

My questioning of the Firebird numbers are due to not having seen Firebird installs, not seeing Firebird people at conventions, and not seeing applications built for Firebird in the OSS space. Heck, I championed inviting Firebird to OSCON and only Ann is attending, as the sole Firebird person at the con. Surely with 1 million users you have 2 people in North America?

I'm getting the sense from the feedback I've seen that FB is used almost exclusively for proprietary software applications, which would tie in with the Interbase history. If that's the case, that would explain why it's mostly passed me by; I've been 100% open source for years.

Comments?

--Josh Berkus
PostgreSQL Project
(my opinions are my own, not the Project's)

Title of the Article
by Josh Berkus on Thu 5th May 2005 16:51 UTC

Paul,

Following up on your post and e-mail:

The reason you got a harsh reaction from a lot of PostgreSQL people was your outright dismissal of PostgreSQL based on the reported installed base. There's the real sense that we're being "punished" in the press for being realistic about the number of running servers instead of making up some inflated number. Since you posted your article on a more interactive site, you get the brunt of that outrage.

The other issue that a *lot* of people have with the article is that it starts out as a comparison of mid-market databases, but the only product covered in any detail is FireBird. Given that MSSQL and MySQL are the two most popular SQL database servers in the world by any measure, that makes the article not a comparison.

Like I said, if the article were titled "Firebird is Popular with Mid-Market" I'd have no problem with it; that seems to be what the article is about. And I can certainly believe that FB is popular in the mid-market, its feature set seems almost perfected targeted at that area.

--Josh Berkus
PostgreSQL Project
(my opinions are my own, not the Project's)

RE:Thanks + Installed base
by Uno Engborg on Thu 5th May 2005 17:11 UTC



I asked about the number for PostgreSQL on the pg-advocacy mailing list, but never received a reply other than that the moderator was reviewing my post. So far, I think the 100,000 number might be in the right ball park as well. My purpose is not to talk down Postgres, so Josh, I would appreciate all help I can get from the PG community with getting numbers for any and all of the systems listed above.


Like you, I have no idea how many Postgresql users there really are. If they only are 100.000, they certainly manage to make big ripples on the Internet.

If you do a search on Postgresql and SQL you get 2 470 000 hits. If you search for MySQL and SQL you get 6 190 000 hits, Oracle and SQL get 7 440 000 hits ,Firebird and SQL get 663 000 hits,

RE:postgresql users
by Uno Engborg on Thu 5th May 2005 17:16 UTC


Paul Richter, I doubt there are only 100,000 postgresql users world wide. After all, apple uses it underneath their desktop nowadays if I understand correctly.


Are you sure about that, I'm quite certain that one of our customers Panther server came with MySQL or do they have postgresql as well hidden away somewhere?

By the does it really matter if its 100.000, 1000.0000 or more. Like Firebird, Postgresql is a very good database regardless of how many actually uses it.

Installed base
by Paul Richter on Thu 5th May 2005 20:47 UTC

I have received several interesting mails about installed base numbers and intend to write a follow-up piece about that.

I think the "downloads x factor" approach is highly inaccurate, to the point of being unusable. The number of pages indexed on Google also is not a measure of real production use.

Real production use numbers show how well a system holds up against real business requirements. It should not count "fun installs" on a developer machine or default installs which are never used.

One route to estimate that is to estimate the installed base of applications. Oracle was pushed ahead by SAP and Oracle ERP apps, DB2 was pushed ahead by AS400 apps, SQLServer/MSDE was pushed ahead by being used in C/S apps built with VB, Firebird/Interbase was pushed ahead by C/S apps built with Delphi/C++Builder, MySQL was pushed ahead by LAMP. These combo's became popular because they really delivered. Sybase lost out against Oracle in the early 90's because its page locking scheme meant it could not run SAP and alike effectively in real world situations.

Which applications have been pushing Postgres ahead? This is not a flame puller, but a sincere question. If we can identify some of the major apps that PG is the common choice for, then perhaps we can get to a substantiated installed base estimate for production use.

Paul

fb and pg...
by Ceaser on Thu 5th May 2005 23:43 UTC

i love firebird, but i just want people to turn away from access, mysql, filemaker, etc to a real db system. either firebird or postgress. mvcc is the bomb! i think they occupy the same user space, but as long as there is no attrition (ie: we loose one or the other). i just want ppl to migrate to either one.

installed userbase doesnt mean much to me.
i'll be very happy when pg doesnt need to be vaccumed, etc (firebird has the same deal more or less).

RE:fb and pg...
by Uno Engborg on Fri 6th May 2005 03:21 UTC


i'll be very happy when pg doesnt need to be vaccumed, etc (firebird has the same deal more or less).


According to the Postgresql TODO list, it will be fixed in the next release. So you will only have to wait a couple of months.