Hello, Just a short reply on Anders speculations of what this object is. This is not an object to move from KO world to BrDb world because that would essentially mean having two databases. It is a root object which gets data from the run db in a way that makes it look like it is getting data out of the BrDb world. It fits into the framework of the BrDb world. Christian just sent out a message telling how to do this more "properly" and this might even merge the BrDb and KO world. The BrDb framework is flexible enough to accomodate this. I have to digest everything he said and try to implement it. But I certainly don't find reason under any circumstances to make a copy of the data from the run database to a different "world". Kris Anders Holm wrote: > Hej Kris, Christian (and Yury too) > > I do not understand in detail what Kris is doing with his BrRun object, > and what Christian is so much against. > > But there is a need to have Run information available when new parameter > values (revissions) are entered into the calibrations data base. > Revissions need validity time information and that comes from the run > information. > > So if Kris is trying to get Run info from the KO-world into the > BrDb-world then that is a good initiativ. What I did not get is if Kris's > BrRun object is a tool to move data from KO-world to BrDb-world or an > object we will all use when we get Run info out of the BrDb-world. > > Right now a consistent SET OF data bases (the Main db and its sub data > bases) is ensured by the utility program(s) that create the data bases. > BrMainDB does not create any data bases. The consistency within a data > base is ensured by using its metods for defining the tables when the data > base is created. > > So if Kris is using a (possibly adjusted) BrRunsDB class with table > definitions to create the tables in a newly created run data base and > ensures that it is registered with the main data base in its DB table so > that it can be connected to via the main db then that must be ok. > > By allways connection via the main db to its sub data bases we ensure that > we work with a consistent SET OF data bases. > > NB: I am right now trying to define and code a set of utility programs for > creation of the data bases with names and hosts as parameters. I assume > that a data base of a given type may be created at a later time than the > main db and then be added to its DB table. We also have to live with > local data bases - at least for testing programs. But I imagine that > there will also be a need to move data between them. > > Anders > > On Thu, 10 Aug 2000, Christian Holm Christensen wrote: > > ? Hi Kris, > ? > ? On Wed, 09 Aug 2000 18:01:22 -0500 hagel@comp.tamu.edu wrote: > ? ? It will be created using BrRunsDb which actually does the Db query to > ? ? the MySQL db on pii3. As I understand things, this BrRunsDb is in turn > ? ? created by BrMainDb (even though I had the courage to do it all by > ? ? myself when I was testing) > ? > ? DON'T DO THAT!!!!!!!!!!!! > ? > ? The reason why the BrMainDb creates the BrRunsDb object is > ? referncial-security!!! BrMainDb looks up the appropiate hostname and > ? db filename, so that you are sure to get the right thing and not have > ? multiple connections, etc.. In general, no object of any class > ? deriving from BrDb should be created by the user, BrMainDb should take > ? care of that. The only exception is BrMainDb, ofcourse, which however > ? is a singleton. > ? > ? The reason why I react so strongly, is because you might encourage > ? others to do the same, with potential disasterours results. So I wish > ? you could keep your dirty tricks to yourself (I do!). > ? > ? Hope you're enjoing the summer, > ? > ? Christian ----------------------------------------------------------- > ? Holm Christensen Phone: (+45) 35 35 96 91 > ? Sankt Hansgade 23, 1. th. Office: (+45) 353 25 305 > ? DK-2200 Copenhagen N Web: www.nbi.dk/~cholm > ? Denmark Email: cholm@nbi.dk > ? > > ========================================================= > Anders Holm email: aholm@nbi.dk > Niels Bohr institute phone: (45) 35 32 52 13 > University of Copenhagen fax: (45) 35 32 50 16 > Blegdamsvej 17 > DK-2100 Copenhagen > Denmark > ========================================================
This archive was generated by hypermail 2b29 : Thu Aug 10 2000 - 12:29:06 EDT