On 26 Apr 99, at 10:14, Alvin Starr wrote:
> On Mon, 26 Apr 1999 luuk@stripped wrote:
> > The competition is maybe a good idea but a lot of database
> > people will say ... no transaction support / subqueries / views
> > you can't get serious about mysql ....
> > that will be another big discussion I think ...
> On the other hand. I could not get Oracle 8.0.5 to run out of the box on
> one of the suported platforms. This left me far less than impressed. After
> I managed to find the problem(Oracle still did not get back to me with a
> fix) I tried to get some support from Oracle to get the Oracle/DBD
> interface software to run on another platform. As you should be able to
> guess I Had no luck. Oracle seems to have a much better marketing program
> than support program.
a bit like microsoft isn't it .... it was for me too hard to get oracle
running so I think we have the same experience on that ... after
some days of looking and trying I had everything working and then
you are asking your self ... this is so slow ... lets start using text
files again if you didn't know anything about mysql ... :-))
> The issue has yet to be settled but I expect our customer will be using
> Mysql instead of Oracle. Why you may ask. The answer is support. There was
> no requirement for functions like transatctions, subqueries or views.
> Granted this is not an extreemly complex application but it is mission
> critical to our customer.
Sorry have no experience with support of oracle so can't say
anything about that ...
> I am not a DBA for a fortune500 company but of the 20-30 projects that I
> have been part of over the last 5 years I have not yet seen a project that
> required the features of Oracle. I also expect that 90-95% of all web
> based applications that require a database can get by quite well with
That's the same way I think about it but don't forget the web isn't
the only place where databases are used and mostly they use the
same db they are using at the office for normal applications. So
maybe we don't have focus on the web part but more on the office
part because their is a lot to win I think.
> > With such a competition you won't say anything about the
> > stability but only about performance.
> It may be possible to collect statistics on patches,bugfixes, and
> workarounds for each DB and then compare the numbers to see who in general
> who has the more reliable product. This of corse will be no easy task.
That will be very hard because not all db companies will post their
bugfixes and patches in the open.
> > Maybe we first have to decide if you can at this moment compare
> > mysql with oracle. Maybe the database engine is faster and more
> > stabel as the one from oracle. But that isn't the only thing
> > oracle is delivering. They deliver also some GUI around it, the
> That is the point of a DB compition. I am sure that there are equivilant
> products to the Oracle products out there that will work with MySQL and a
> hand full of other databses.
> > name, the company etc etc .... and what you can't forget it's
> > easier to find an new employee who knows oracle as one who knows
> > mysql (by thinking of the companies perspective) and I think a
> > lot of people will select the database more on the latest part
> > as of the database engine.
> I hate this argument. It has been used mostly do justify stupid decisions.
It's not justifying stupid decisions but saying what mostly is done.
You don't hear me say that you should oracle instead of mysql ...
the opposite is true ...
> If this argument were really true than we would all be using IBM hardware
> and IBM software to the exclusion of all else. MySQL has a place in the
> range of database products that need to be available to allow people to
> build applications. The thing is to identify the place wher MySQL best
> fits and then use that informatin.
That's true. I think we can say mysql is the best web db their is but
we now have to see how to broader that path to office db's or data
> > But if we would like to start an contest why aren't we trying to
> > do a tcp d benchmark first to see how mysql is doing.
> I would love to have a list of statistics and comparisons that I can point
> to that will help me convince customers that they do not need to use
> Oracle for a 10000 record database that is part of there web presence.
I think for allmost 2 years I am running and developing the
benchmark of mysql (which is included in the source distribution).
All the results of my benchmark test on all the different databases
can you see at the mysql site. (http://www.tcx.se/benchmark.html)
So there is some comparison to show to the customers. The only
problem is is that I am also configuring all the other db's and lately
I managed to get some better performance out of the other db's but
I am absolutely no expert on every db. That's the advantage of the
tpc d benchmark where everybody can run their own benchmark
test and you can share the results to see how they perform to each
other. The only problem is is that mysql doesn't support all the
SQL language that's needed to run the official tpc d benchmark. I
have written a test tpc d benchmark where I changed the queries
so mysql could run them and they show some nice figures but a
pitty it's not compareble to the other run's on other databases.
I think the only way to get a bigger name in database world is to
help monty and tcx to develop mysql any further and we could start
this discussion again over an half year or year again. I am not
saying that mysql isn't a good product at this moment but I
personally think we first have to implement some features to be
real competitive with the big names on a broader range.
| Luuk de Boer -> <luuk@stripped> |
| When you want to go fast ..... |
| Drive a cobra on the road and take a T3 line for internet ...... :-) |
| http://www.luuk.com/ -> The AC Cobra site |