🚀 go-pugleaf

In article <48sptv$9b@senator-bedfellow.MIT.EDU> templon@studbolt.mit.edu (Jeff Templon) writes:
>Path: nntp.teleport.com!gasco!qiclab.scn.rain.com!orcalink.com!imci2!newsfeed.internetmci.com!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!studbolt!templon
>From: templon@studbolt.mit.edu (Jeff Templon)
>Newsgroups: comp.lang.python,comp.databases,comp.databases.object,sci.physics,comp.lang.fortran,cern.general,cern.hepdb
>Subject: [??] seeking apparently unique database/browser package
>Date: 21 Nov 1995 15:06:39 GMT
>Organization: MIT Lab for Nuclear Science, Nuclear Interactions Group
>Lines: 73
>Message-ID: <48sptv$9b@senator-bedfellow.MIT.EDU>
>NNTP-Posting-Host: studbolt.mit.edu
>X-Newsreader: TIN [version 1.2 PL2]
>Xref: nntp.teleport.com comp.lang.python:6823 comp.databases:48701 comp.databases.object:7959 sci.physics:148271 comp.lang.fortran:33105



>Hi,

>I am looking for a very simple database package/browser, so simple
>that all of the messages I've seen in the various relevant newsgroups
>describe products that are much too complicated for what I want
>to do.  Let me describe what I need, and I hope that someone
>out there has seen or written something close to it.

>I am evaluating options for a database browser which needs to
>keep track of experimental or experimentally-determined parameters.
>The database itself can be extremely simple; in fact at the moment,
>we are using a simple ASCII file to hold all these parameters.

>My ideal database and browser would meet the following criterion:

>1) it must be able to save the actual database in some format
>   which could be simply read by a Fortran program.  This program
>   is the primary end-user of the database.  It needs all these
>   parameters as it analyzes experimental data and converts these
>   data to physical observables.  The parameters in the database
>   determine the transformation from data->observable.  An ASCII
>   file with some predictable format, or a Fortran API, is
>   sufficient.
It is not clear to me what you mean.  You are talking about parameters and a 
database.  Do you mean that you have a time series of data, such as voltages 
at specified intervals (for example) and you want to do some mathmatical 
operation (in Fortran) to these observations and produce results.  Quite a few 
commercial databases support a Fortran API.  I know Oracle does, and I am sure 
there are others.  

>2) The browser allows users to do some flexible searching on a 
parameter>   name, so that they do not have to navigate an interminable menu 
system.
It looks like you want some sort of tool to look at data.  There are a zillion 
out there.  You might look at Crystal Reports.  People in our office have 
found it easy to use.  There are others.  It depends upon the level of 
computer sophistication of the users.  I like Crystal for some things, but 
sometimes I want more direct contact with the database.  So by computer 
sophistication I DO NOT mean intelligence; I mean confort with the tool.  Of 
course, Crystal only runs on MS Windows.  

>3) some sort of object-orientation is nice.  there are some sets of
>   parameters that naturally go together, such as the coefficients
>   of transfer matrices which determine the optical properties of
>   a spectrometer.  I would like to be able to move some set of
>   associated parameters in and out of the database as a group.
>   This means that one would like the capability of having a subset
>   of the database also be a valid database from the point of view of
>   the browser; I could copy such a set out of one database (being
>   used, say, by a calibration program) and then import it into
>   another database (that second DB containing lots of other
>   parameters which have already been determined by other means.)

You can create views which means that you can have "virtual" tables (don't 
shoot me for the term) so that you can group things together in quite a number 
of different ways.  The advantage of a view is that you do not have to move or 
copy data from one place to another and thus have to worry about it getting 
stale.  Most commercial RDBMS's databases support this.

>4) some facility for adding comments to the database would be
>   nice, so that the database would be to some extent self-documenting.
Most commercial RDBMS's databases support this. Usually via the comment on 
command

>5) A flexible means of adding fields to the database and modifying the
>   grouping structure.  An embedded scripting or extension language
>   might also come in handy; you never know what you'd want to
>   do in the future.
Most databases allow the adding of fields.  I do not know what grouping 
structure means.  If it means I want these attributes for this thing then 
most databases do this.  Quite a few databases have an extension language.  
Of course, the extension language is very database oriented.  I would not 
try to do some of the things you do in  Fortran in the database.  It might 
be possible, but a real pain in the neck.  Easier to do it in Fortran and 
just retrieve the data you want.  If you do not know what you want to do then 
you cannot find anyone to furfill this requirement.  Sort of like: "Mr. 
architech I want a house that is easy to change it into something else at some 
future date."

>6) As far as I know, network access is not a big deal (many of the
>   database posts I see seem to be associated with lots of people
>   using a database simultaneously over some network.)  Performance
>   also should not be a big deal, since the database will be read
>   in only once per instance of the program.  We would like to
>   have reasonable performance for starting up the browser, since
>   it might get killed and restarted frequently; so say startup
>   times less than ten seconds.
It sounds like you have very little data.  If you are reading in all the data 
and the total startup time is less than 10 seconds then you cannot have much 
data.  That is maybe 20  megs of data.  I mean it does take time to 
read stuff off disk.  So if it takes 1 second for the program and 9 seconds 
max for the data and at 2 megs per second that means 18 megs. Not a lot of 
data for a commercial database.  

>I can see how one could implement such a system fairly easily using
>a text-file for the DB, Python for the language (or pick your
>favorite language), and Tk for the graphical interface, but it'd
>be a lot of work, and I can't believe that noone has ever made
>something like this!  Please, if you know of something like this,
>send me an email so we can discuss it.

>                                                JT
>   
You state that it would be a lot of work.  I am sure you are right.  
Jim Kennedy

>The "lapel pin" [xx] in the Subject of this message signals
>     adherence to SELF-DISCIPLINE, a voluntary program for
>     quality News discussions. For details see
>     http://www.eiffel.com/discipline or
>     ftp://ftp.eiffel.com/discipline. Please do NOT post
>     comments on this newsgroup on this aspect of the message.