Thread: Survey results from the PostgreSQL portal page
Hi everyone, Dave Page put up a new survey on the PostgreSQL portal page very recently, " What would attract the most new PostgreSQL users?" and the results in already are interesting (1,529 results as this is being written): http://www.postgresql.org/survey.php?SurveyID=9 Listed from most voted for to least voted for we have: *********** Answer Responses Percentage More speed 505 33.028% Win32 Port 390 25.507% Replication 386 25.245% Better docs 160 10.464% More features 32 2.093% Better marketing 29 1.897% Better migration 18 1.177% PITR 9 0.589% Total number of responses: 1529 *********** Now, we don't necessarily have a speed problem, as people who take the time to tune the database can attest to, so this is making me consider why such a large percentage of folk would vote for that. The possibilities that come to mind immediately are: + People don't know that they should tune the database, and are leaving the configuration settings at the defaults. We could adjust the perception of PostgreSQL's speed for these people by adjusting the default settings. We were already considering raising the memory buffer defaults weren't we? + People are having troubles related to VACUUM. This is being worked on presently isn't it? + People don't know *how* to tune the database properly yet? + Maybe we need more inbuilt self-tuning abilities or utilities for PostgreSQL? Other interesting conclusions can be drawn from the results too, one of which is that only about 2% of people are asking for more features, and also that only about 2% are looking for better marketing. Anyway, thought this worth bringing to people's attention, as we may find some value in it. :-) Regards and best wishes, Justin Clift -- "My grandfather once told me that there are two kinds of people: those who work and those who take the credit. He told me to try to be in the first group; there was less competition there." - Indira Gandhi
On Sunday 19 January 2003 09:20, Justin Clift wrote: > Dave Page put up a new survey on the PostgreSQL portal page very > recently, " What would attract the most new PostgreSQL users?" and the > results in already are interesting (1,529 results as this is being > written): > > http://www.postgresql.org/survey.php?SurveyID=9 > > Listed from most voted for to least voted for we have: > > *********** > > Answer Responses Percentage > > More speed 505 33.028% > Win32 Port 390 25.507% > Replication 386 25.245% > Better docs 160 10.464% > More features 32 2.093% > Better marketing 29 1.897% > Better migration 18 1.177% > PITR 9 0.589% > > Total number of responses: 1529 What I find interesting is that 25% voted for replication and only 1/2% voted for PITR. I think that that shows that surveys are easily skewed by their own parameters. People interested in both probably just voted for the one slightly higher on their wish list. Anyway, interesting straw vote but let's not make critical decisions based on it. It's not even a vote of the people looking for databases. It's a vote of what those people want in the opinion of the people who voted. > Now, we don't necessarily have a speed problem, as people who take the > time to tune the database can attest to, so this is making me consider > why such a large percentage of folk would vote for that. Need for speed. It doesn't matter how fast we are, faster is always better. Other items are of the nature have or not have a feature and people may disagree whether we need the feature or not but who will ever say that more speed is not wanted? -- D'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves http://www.druid.net/darcy/ | and a sheep voting on +1 416 425 1212 (DoD#0082) (eNTP) | what's for dinner.
On Sun, 2003-01-19 at 15:20, Justin Clift wrote: > > Now, we don't necessarily have a speed problem, as people who take the > time to tune the database can attest to, so this is making me consider > why such a large percentage of folk would vote for that. > > The possibilities that come to mind immediately are: + people measure postgresql by the speed of bulk imports which seem to be faster on other dbs (I'm only repeating what I often see on the lists), and don't consider that this is not something most people do on a regular basis. btw, PITR would get more hits if more people knew what it means... (Yes, I know what it is, but I think many who don't read the mailing lists etc. don't). cheers -- vbi -- featured link: http://fortytwo.ch/gpg/subkeys
On Sun, 2003-01-19 at 14:20, Justin Clift wrote: > Dave Page put up a new survey on the PostgreSQL portal page very > recently, " What would attract the most new PostgreSQL users?" ... > Other interesting conclusions can be drawn from the results too, one of > which is that only about 2% of people are asking for more features, and > also that only about 2% are looking for better marketing. I suspect the majority of those who responded are technical people who despise marketing. I also suspect that most people didn't answer the question asked but instead said what they themselves most wanted. But the only thing that will get many more users is much better marketing. If people don't hear about PostgreSQL, they will never even think of using it. I looked at the shelves of database books in Blackwells in Oxford yesterday: lots on Oracle and Sql Server and DB and several on MySQL. There were 2 on Postgresql, and only one copy of each. I'd be interested to know what the commercial PostgreSQL companies think about it -- Oliver Elphick Oliver.Elphick@lfix.co.uk Isle of Wight, UK http://www.lfix.co.uk/oliver GPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839 932A 614D 4C34 3E1D 0C1C ======================================== "The LORD is my strength and song, and he is become my salvation; he is myGod, and I will prepare him an habitation; my father's God, and I will exalt him." Exodus15:2
> > >+ people measure postgresql by the speed of bulk imports > > This is a good point. I can complete agree. What we might need is something called "SQL Loader" or so. This may sound funny and it doesn't make technical sense but it is an OBVIOUS way of importing data. People often forget to use transactions or don't know about COPY. A perfectly documented tool called "SQL Loader" or so might help. This is mainly a marketing reason but according to the things I have seen this is true. + people think about MySQL and speed People and magazines have MySQL in mind and this is a point which annoys me most. Why don't they just think about SAP DB or Oracle instead of MontySQL. >which seem to be faster on other dbs (I'm only repeating what I often >see on the lists), and don't consider that this is not something most >people do on a regular basis. > >btw, PITR would get more hits if more people knew what it means... (Yes, >I know what it is, but I think many who don't read the mailing lists >etc. don't). > > People with more experience will ask for more high level features. People need to get started quickly so that they will find out about the advantages of a database. Features have to be obvious because other way people with not much experience won't look for it. Also: At least once a week somebody is asking for replication. Replication, clustering, hot failover, and so forth are definitely the features which are asked by companies. I wonder why people ask for better documentation. I think the documentation is really good. Ever read Oracle stuff? *ugh*. Hans -- *Cybertec Geschwinde u Schoenig* Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/1/913 68 09; +43/664/233 90 75 www.postgresql.at <http://www.postgresql.at>, cluster.postgresql.at <http://cluster.postgresql.at>, www.cybertec.at <http://www.cybertec.at>, kernel.cybertec.at <http://kernel.cybertec.at>
Justin Clift <justin@postgresql.org> writes:
> Dave Page put up a new survey on the PostgreSQL portal page very
> recently, " What would attract the most new PostgreSQL users?" and the
> results in already are interesting (1,529 results as this is being written):
> [snip]
> Now, we don't necessarily have a speed problem, as people who take the
> time to tune the database can attest to, so this is making me consider
> why such a large percentage of folk would vote for that.
Interesting question. Too bad the survey didn't ask *what* they find
slow.
> Other interesting conclusions can be drawn from the results too, one of
> which is that only about 2% of people are asking for more features, and
> also that only about 2% are looking for better marketing.
Of course, this is a survey of people who already know about Postgres,
and are sufficiently interested in using it to have visited the website.
So really, people who need to be marketed to weren't surveyed. The low
vote for that point is probably just survey skew.
regards, tom lane
On Sun, Jan 19, 2003 at 09:43:03AM -0500, D'Arcy J.M. Cain wrote: > What I find interesting is that 25% voted for replication and only 1/2% voted > for PITR. I think that that shows that surveys are easily skewed by their > own parameters. People interested in both probably just voted for the one > slightly higher on their wish list. This one surprised me too. Almost all companies I talked to abolut PostgreSQL ask for PITR. Yes, many also ask for replication, but the amount is much lower. Michael -- Michael Meskes Email: Michael@Fam-Meskes.De ICQ: 179140304 Go SF 49ers! Go Rhein Fire! Use Debian GNU/Linux! Use PostgreSQL!
> I wonder why people ask for better documentation. I think the > documentation is really good. Ever read Oracle stuff? *ugh*. They want examples of real-world usage. The commands themselves have good 'HOW TO' notes, and an explanation of what they are, but we don't really have anything on 'WHY?'. A large number of people coming from MySQL, or just starting out don't know what feature XYZ can do for them, and why it's best (easiest) for the database to do that work -- regardless of what that Java programmer thinks beans are for. Much of this can probably be borrowed from books on basics. -- Rod Taylor <rbt@rbt.ca> PGP Key: http://www.rbt.ca/rbtpub.asc
> On Sunday 19 January 2003 09:20, Justin Clift wrote: >> Dave Page put up a new survey on the PostgreSQL portal page very >> recently, " What would attract the most new PostgreSQL users?" and the >> results in already are interesting (1,529 results as this is being >> written): >> *********** >> >> Answer Responses Percentage >> >> More speed 505 33.028% >> Win32 Port 390 25.507% >> Replication 386 25.245% >> Better docs 160 10.464% >> More features 32 2.093% >> Better marketing 29 1.897% >> Better migration 18 1.177% >> PITR 9 0.589% >> >> Total number of responses: 1529 > The funny thing is that, to me, this least is perfectly reflective of the FUD that other databases use against postgresql. It's too slow, it only runs on *nix, it doesn't have replication, and the documentation isn't very good. You can't FUD postgresql on feature set, becuase we have a pretty wide feature set (as good as any other open source rdbms afaik) plus it's open source, so if we don't have a feature that say oracle has, you can pay someone the $10,000+ the oracle license will cost to implement it. I've also not seen much FUD on the other issues either. If you can address the primary points that your competition is using as FUD, you will gain new users. We'll see what happens in 7.4 if we do have replication, native windows support, and PITR, because everyone will have to come up with some new FUD to sling this way. Robert Treat
> -----Original Message----- > From: Adrian 'Dagurashibanipal' von Bidder > [mailto:avbidder@fortytwo.ch] > Sent: 19 January 2003 14:47 > To: PostgreSQL Hackers Mailing List > Subject: Re: [HACKERS] Survey results from the PostgreSQL portal page > > > btw, PITR would get more hits if more people knew what it > means... (Yes, I know what it is, but I think many who don't > read the mailing lists etc. don't). I've fixed that now. It ain't too pretty but it should appear after the next hourly site rebuild. Regards, Dave.
On Sun, Jan 19, 2003 at 01:19:03PM -0500, Robert Treat wrote: > pretty wide feature set (as good as any other open source rdbms afaik) > plus it's open source, so if we don't have a feature that say oracle has, > you can pay someone the $10,000+ the oracle license will cost to implement > it. I've also not seen much FUD on the other issues either. If you can Unfortunately it doesn't always work this way. I knew one government organization that decided to go for Oracle for 500K Euro instead of adding the missing features (actually almost exclusively PITR). One of the top arguments I heard was: "I don't believe that free software community works. Once the developers get a social life or even kids, they stop working on software." Of course I told him that I still do work on free software despite having three sons on which he answered: "Maybe, but I still don't believe it." Sad but true. Michael -- Michael Meskes Email: Michael@Fam-Meskes.De ICQ: 179140304 Go SF 49ers! Go Rhein Fire! Use Debian GNU/Linux! Use PostgreSQL!
Michael Meskes <meskes@postgresql.org> wrote: > Unfortunately it doesn't always work this way. I knew one government > organization that decided to go for Oracle for 500K Euro instead of > adding the missing features (actually almost exclusively PITR). One of > the top arguments I heard was: "I don't believe that free software > community works. Once the developers get a social life or even kids, > they stop working on software." Of course I told him that I still do > work on free software despite having three sons on which he answered: > "Maybe, but I still don't believe it." > > Sad but true. I think it is about "sustainability of free software". Linux kernel community gets many contributions and commitments from big IT companies, like Intel, IBM, NEC, etc. They throw many hackers into opensource community. So I think the linux kernel is getting more powerful and sustainable. Linux kernel can be trustful in the future? Yes, I think so. Even today, Oracle is running on Linux. In opensource community, I think only the developer pool can be guarantee for its sustainability. We need more developers and supporting companies. Can we get a belief in sustainability? BTW, Japanese government is considering the opensouce software. One official said, "We don't want to pay more license fee to the proprietary software companies, like MS, Oracle or else. But we don't know how to keep the relationship with opensource community. The government may tend to be disliked by NPO(Non-profit organization) or else...". But he also said, "NAIST(www.naist.go.jp) are going to replace *ALL* their software with opensource in next 2 years." I feel we are in the new movement now. We have a chance in our hands. Don't be disappointed. And my project is granted by IPA(www.ipa.go.jp) Japan. :-) -- NAGAYASU Satoshi <snaga@snaga.org>
Tom Lane wrote: > Justin Clift <justin@postgresql.org> writes: > >>Dave Page put up a new survey on the PostgreSQL portal page very >>recently, " What would attract the most new PostgreSQL users?" and the >>results in already are interesting (1,529 results as this is being written): >>[snip] >>Now, we don't necessarily have a speed problem, as people who take the >>time to tune the database can attest to, so this is making me consider >>why such a large percentage of folk would vote for that. > > > Interesting question. Too bad the survey didn't ask *what* they find > slow. Hey, interesting point. Perhaps for the next survey, we should ask which part of PostgreSQL people feel needs the greatest speed increase? We just have to think of the best wording for it... :) >>Other interesting conclusions can be drawn from the results too, one of >>which is that only about 2% of people are asking for more features, and >>also that only about 2% are looking for better marketing. > > > Of course, this is a survey of people who already know about Postgres, > and are sufficiently interested in using it to have visited the website. > So really, people who need to be marketed to weren't surveyed. The low > vote for that point is probably just survey skew. Yep. It's a reasonably ad-hoc system, but it does provide some interesting aggregate feedback. It's nice that we receive a decent amount of votes over a short amount of time. Out of curiosity, any thoughts as to what other questions or topics might be good to be asked about? :-) Regards and best wishes, Justin Clift > regards, tom lane -- "My grandfather once told me that there are two kinds of people: those who work and those who take the credit. He told me to try to be in the first group; there was less competition there." - Indira Gandhi
> I wonder why people ask for better documentation. I think the > documentation is really good. Ever read Oracle stuff? *ugh*. Ever read MySQL docs - *hack*!! Chris
Michael Meskes wrote: > On Sun, Jan 19, 2003 at 01:19:03PM -0500, Robert Treat wrote: > >>pretty wide feature set (as good as any other open source rdbms afaik) >>plus it's open source, so if we don't have a feature that say oracle has, >>you can pay someone the $10,000+ the oracle license will cost to implement >>it. I've also not seen much FUD on the other issues either. If you can > > > Unfortunately it doesn't always work this way. I knew one government > organization that decided to go for Oracle for 500K Euro instead of > adding the missing features (actually almost exclusively PITR). One of > the top arguments I heard was: "I don't believe that free software > community works. Once the developers get a social life or even kids, > they stop working on software." Of course I told him that I still do > work on free software despite having three sons on which he answered: > "Maybe, but I still don't believe it." > > Sad but true. Interesting observation, and not entirely irrelevant. It's the strength of any particular Open Source Community that seems to indicate whether or not there are going to be enough people getting involved to overcome the attrition rate of the people becoming less involved. With PostgreSQL, a lot of work goes into building and feeding the community. That includes making sure the right people are talking to each other, assisting people to find the information they need, and other simpler stuff like making sure the basic facilities work (cvs, ftp, websites, etc). We are fortunate in that being based on a BSD license is assisting businesses to adopt PostgreSQL without needing to think too hard about licensing ramifications, and we are also fortunate that the quality of PostgreSQL is extremely good and has an increasingly excellent reputation that is attracting people from countries all over the world to get involved. When people suggest that the "Free Software Community" doesn't work, it may be worthwhile pointing out that it works very well for the Communities that are strong, but he could be correct for those that haven't become self-sufficient yet. :-) Regards and best wishes, Justin Clift > Michael -- "My grandfather once told me that there are two kinds of people: those who work and those who take the credit. He told me to try to be in the first group; there was less competition there." - Indira Gandhi
On Mon, 20 Jan 2003, Christopher Kings-Lynne wrote: > > I wonder why people ask for better documentation. I think the > > documentation is really good. Ever read Oracle stuff? *ugh*. > > Ever read MySQL docs - *hack*!! The documentation definately needs work -- particularly client library documentation and PL/PgSQL. I want to work on this when I get time. Gavin
Gavin Sherry wrote:
> On Mon, 20 Jan 2003, Christopher Kings-Lynne wrote:
>
>
>>>I wonder why people ask for better documentation. I think the
>>>documentation is really good. Ever read Oracle stuff? *ugh*.
>>
>>Ever read MySQL docs - *hack*!!
>
>
> The documentation definately needs work -- particularly client
> library documentation and PL/PgSQL. I want to work on this when I get
> time.
Case in point : in 7.3, the ODBC driver documentation (which was terse and
somewhat outdated, to begin with ...) has disappeared from the main tree.
You have to go to GBorg to find (some) relevant information (and no
examples, BTW). But to find an information I really needed, I had to use
... the driver source, fer Crissakes !! I felt back in '74, when I tried to
learn Fortran.
[ BTW : note to Hiroshi Inoue : Thank you ! I partially solved by problem,
and think a real solution is bound to to undoing some 7.2 to 7.3
modifications ...]
The same could be said of the JDBC driver, btw, while it's doc is still in
the main doc tree.
This one is one of my pet peeves at the moment ...
Emmanuel Charpentier
On Sun, 19 Jan 2003, [ISO-8859-1] Hans-J\xFCrgen Sch\xF6nig wrote: > >+ people measure postgresql by the speed of bulk imports > > This is a good point. I can complete agree. What we might need is > something called "SQL Loader" or so. This may sound funny and it doesn't > make technical sense but it is an OBVIOUS way of importing data. People > often forget to use transactions or don't know about COPY. Even "doing it right," postgres 7.2 was significantly slower than MySQL for bulk data imports, at least for tables with relatively narrow rows. I was going to put this down to higher row overhead, except that it was nowhere near raw file I/O speed, either. So this could use improvement, if it's not been improved already. There's room for performance increases in a lot of other areas, too, but in the end, a lot of people just don't design their databases for good performance. And I've killed enough non-postgres database servers in my life to know that if you don't really know what you're doing, you can easily make the performance of any DBMS suck. :-) Personally, I think there's still a fair amount of room in the features area, too. I'm always running into something that I'd like to have. Today it was being able to defer a UNIQUE constraint. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're alllight. --XTC
Michael Meskes wrote: > On Sun, Jan 19, 2003 at 01:19:03PM -0500, Robert Treat wrote: > > pretty wide feature set (as good as any other open source rdbms afaik) > > plus it's open source, so if we don't have a feature that say oracle has, > > you can pay someone the $10,000+ the oracle license will cost to implement > > it. I've also not seen much FUD on the other issues either. If you can > > Unfortunately it doesn't always work this way. I knew one government > organization that decided to go for Oracle for 500K Euro instead of > adding the missing features (actually almost exclusively PITR). One of > the top arguments I heard was: "I don't believe that free software > community works. Once the developers get a social life or even kids, > they stop working on software." Of course I told him that I still do > work on free software despite having three sons on which he answered: > "Maybe, but I still don't believe it." > > Sad but true. One issue he is probably right about is that more burden is placed on the user for testing/support in open source than in closed source. Of course, open source is usually free, so you can afford to pay for those extras, but they do exist have have to be managed. I bet some companies just want to pay the bill and the yearly support and don't want to deal with the extra burden, even if it saves them money. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073