Re: Quote from student, after teaching Pick
- From: "Excalibur" <excalibur21@xxxxxxxxxxx>
- Date: Thu, 17 May 2007 23:52:04 GMT
Hi
I firmly agree about the language simplicity issue.
I am afraid that I have to disagree on some points.
I ventured into the RDBM world and could not get back quick enough on the
very point of database design. The simplicity of change was a major point.
The ability to use ENGLISH/ACCESS/INFO etc to achieve quite complex results
in a couple of minutes when the user changed his mind 6 months after going
live, was a second issue. Originally speed was a third although now less
relevant.
Resilience has not been an issue since the introduction of the FSI. I still
insist on a UPS and build in validation checks but overall the D3 FSI on
Windows is terrific.
One thing that is required in Pick is programmer discipline. All the issues
I have seen over the years have come from sloppy work. Unfortunately it
appears from recent posts from Anthony and Usha that some people think that
a strict database ,separate from the logical area, especially when combined
with objects done by experts can solve this problem. I am afraid that they
are very wrong. A classic I ran over was the many months that it took SQL
"experts" to get cheques printing at better than one a minute from the
Victorian Education Dept payroll. Finally a genuine expert waltzed in at
9am and left for lunch with the things flying off the printer.
Peter McMurray
"mike ryder" <mg.ryder@xxxxxxxxx> wrote in message
news:1179400824.765049.104340@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Interesting discussion, but IMHO, flawed logic.
1) People like the PICK model for its simplicity of language not its
design of database. If there were a product available that enabled the
language but talked to some other database (with the same performance)
then everybody would dance for joy. jBase, Ondata and Revelation
appear to provide that step.
2) Resiliance - who would buy a database that gets a corruption
because the server power failed? Big database restores - yuk!! ;-)
3) Multivalued - both Oracle and DB2 can store data in xml structures
(which has more multivalues than you could possibly conceive), but the
storage model for each is completely different.
Oracle stores the data in a method that requires a disk read for every
record, but, by using structured storage, you can get Oracle to use
the buffers. Using structured storage requires that the data for each
field fits a data type and size - bit of a problem for most PICK
applications
DB2 understands xml natively and can retrieve a single record or
attribute (field) (in the PICK sense) and it can use buffers to store
this level of data. DB2-9 is probably faster than a native PICK
database in reading / writing multivalue data.
It would, therefore, seem likely that the multidimensional model is
coming into flavour for all of the reasons that we have all loved
PICK, but it is unlikely to be manifested as a boom in PICK. It is
dangerous to assume other databases are also remaining static as is
the PICK market.
Just my 2c
Mike
.
- References:
- Quote from student, after teaching Pick
- From: dawn
- Re: Quote from student, after teaching Pick
- From: Anthony Lauder
- Re: Quote from student, after teaching Pick
- From: Tony Gravagno
- Re: Quote from student, after teaching Pick
- From: dawn
- Re: Quote from student, after teaching Pick
- From: Glen B
- Re: Quote from student, after teaching Pick
- From: dawn
- Re: Quote from student, after teaching Pick
- From: Glen B
- Re: Quote from student, after teaching Pick
- From: Jeff Caspari
- Re: Quote from student, after teaching Pick
- From: Glen B
- Re: Quote from student, after teaching Pick
- From: mike ryder
- Quote from student, after teaching Pick
- Prev by Date: Re: Quote from student, after teaching Pick
- Next by Date: Re: Quote from student, after teaching Pick
- Previous by thread: Re: Quote from student, after teaching Pick
- Next by thread: Re: Quote from student, after teaching Pick
- Index(es):
Relevant Pages
|
Loading