Re: database market share 2003

Its pathetic that you can't write good code. ;-) If you can grok C++, then you should be able to write thread safe C code and then use them. And actually if you're wroting in C++ then you need to have your head examined. But thats a whole different flame war. ;-) But I bet you have trouble with JDBC and how J2EE tries to use the database only as a means to make objects persistent. ;-) But hey what do I know? ;-) I"m just an old school programmer. Jean-David Beyer wrote: > Daniel Morgan wrote: > >> Mark A wrote: >> >>> "robert" <gnuoytr@rcn.com> wrote in message >>> news:da3c2186.0405291715.5e4125bd@posting.google.com... >>> >>>> what i've not seen is the other side of this coin: that (perhaps) >>>> IBM's share is skewed (looks larger than it really is) by the >>>> fact that it pretty much owns the mainframe. a relative handful >>>> of very expensive installs. in other words, i question how relevant >>>> DB2 is to the future of relational databases. IBM needs to >>>> demonstrate that it is relevant outside of conversions (i use >>>> the term very, very loosely) of behemouth COBOL/VSAM systems. at >>>> my work, they just defined tables from the copybooks. i >>>> gather this is quite common. >>>> >>>> robert >>> >>> >>> >>> >>> Since DB2 mainframe has been around since the mid-1980's, that is >>> ridiculous. The overwhelming majority or DB2 OS/390 applications were >>> designed on DB2 from scratch. Your company may be an exception, and >>> somewhat >>> backward. After all, they employ you, so it must be a really screwed up >>> company. >> >> >> >> But for how much longer ... I wonder? >> >> I am watching the huge inroads being made by clustered Linux taking out >> Sun's and H/P's more expensive offerings. I built an 8 CPU cluster a few >> weeks back with less than $11,000 US in hardware. >> >> How long before it becomes easy to build OS/390 equivalent machines >> with a rack of 2 CPU x 4GB Intel boxes running RedHat AS? >> >> I suspect far sooner than you want to imagine. >> >> And when the big iron goes ... do you think DB2 will survive? Informix, >> in my opinion, has a better chance of surviving. >> > Well, I ran Informix on my Red Hat Linux 5.0 box for a while. It was a > bit pathetic in that it did not have an API for C++: only for C. So I > ended up writing a large bunch of C functions to interface to the > Informix server. It worked, but sloppy. Then I upgraded to Red Hat Linux > 6.0 and Informix never worked again. I e-mailed Informix about it and > they said they did not know if it would work with 6.0 or not. I said I > would be glad to test it for them, but that the CD-ROM I had would not > read and could they send me another one. They never answered any e-mails > after that. > > Which is why I upgraded to IBM DB2 UDB V6.1 which worked just fine. > Since I am cheap, I kept V6.1 through upgrades to RHL 6.2 and 6.3. Now > DB2 V6.1 did not really like RHL 7.3. It worked because there were some > "compatability libraries" that could be used, but it meant I had to > change all the makefiles to use those libraries, including getting it to > use the compatibility version of ldd.so. > > Anyhow, when I built this machine I put Red Hat Enterprise Linux 3 ES on > it and installed IBM DB2 UDB V8.1 (upgraded to 8.1.5) and after a fight > to get it installed (graphic installer does not work), it runs just fine > on a dual hyperthreaded Intel XEON machine with 4GB (expandable to 16 GB > if I found I needed it). > > Unless Informix has been greatly improved since the time RHL 6 came out, > I do not see why anyone would wish to use it unless it is a lot cheaper > than DB2. > > My needs are quite modest, since I am running it single-user for a > single small (by dbms standards) database. But back when I started, > postgreSQL and Oracle were the other alternatives, and postgreSQL did > not run right (one version would not allow primary keys to be specified, > and another was unable to have views), and Oracle's license agreement > was so complicated that I refused to sign it. > > If a small user such as I can use it, I see no reason why a larger user > could not. >

nobody wrote: > Its pathetic that you can't write good code. ;-) It's pathetic that you top post. ;-) > > If you can grok C++, then you should be able to write thread safe C code > and then use them. I wrote C code for 25 years or so (since early 1970s). I have not needed to write thread-safe code, since I find that Linux creates and destroys processes cheaply enough that I just run multiple processes instead of threads. If I need to share memory, I do so explicitly. > And actually if you're wroting in C++ then you need > to have your head examined. But thats a whole different flame war. ;-) And if you think _wroting_ is an English word, you need to spend time in a re-education camp. This is 2004, and it is pretty boring to engage in that tired old flame war. > > But I bet you have trouble with JDBC and how J2EE tries to use the > database only as a means to make objects persistent. ;-) > > But hey what do I know? ;-) > I"m just an old school programmer. > Bah! I am an older school programmer thay you, probably. I have programmed in three different assemblers for the IBM 704 alone, not counting NY-AP1. -- .~. Jean-David Beyer Registered Linux User 85642. /V\\ Registered Machine 241939. /( )\\ Shrewsbury, New Jersey http://counter.li.org ^^-^^ 14:55:00 up 1 day, 20 min, 4 users, load average: 2.26, 2.15, 2.10

Jean-David Beyer wrote: > nobody wrote: >> Its pathetic that you can't write good code. ;-) > > It's pathetic that you top post. ;-) >> >> If you can grok C++, then you should be able to write thread safe C >> code and then use them. Well, this flame war was unexpected. Do people still disagree that rkuserthingy is a dickhead for starting this thing?