Wish list? or exists? 
Author Message
 Wish list? or exists?

I have a wish list, and since there is nowhere on Borland On-Line to
deposit it I think I'll drop it here.

If anyone knows of a fix, please let me know, else it will remain,
as it was intended to be, a wish list.

(1) Lookup Tables

It seems that "relocating" a database and/or form within
Paradox can be an experience in itself.  Hardcoded lookup
tables (ie drop down edit lists) that make reference to
Aliases seem to have no difficulty when a form and/or
database is relocated to another drive/directory.
However, lookup tables defined in the "restructure" procedure,
although the alias may be used to define the location
of the lookup table, Paradox seems to code the PHYSICAL location
of the lookup table rather than the ALIAS path to the table
as specified in the IDAPI configuration utility.

This holdback is extremely frustrating.  Our clients directory
structure will most certainly differ from the structure of
our office network drive.  This limitation could lead to many
problems if an application's directory structure is modified in
any way.

Also, it has occured to me that when a database cannot located
it's defined lookup a dialog box appears prompting the user
to pick from 3 options, Fail Lookup, Ignore, or Remove Lookup.
This feature does not seem to function properly.  In many
instances, none of these choices will allow the user to access the
table in any fashion.

Using the table repair utility, and modifying the structure of the
table within the utility also has no effect.  The new table
created by the utility has the same problem as the unmodified table.

(2) Multi-level referential integrity and multi level cascading.

The ability to link a master table to several sub-levels of child tables
is not possible withing Paradox 5.0.  This seems to be an inhibitor
to the development of multi-level database structure.

Referential integrity should also be able to be established between
tables in other directories and not limited to tables within
the master only.

A child table, capable of referencing a table which subsequently
references another master would be an attractive item.

Any comments or any more wishes??

Mail me

Anthony V. Parente



Sun, 28 Dec 1997 03:00:00 GMT
 Wish list? or exists?


---------->  text deleted  <--------

Quote:
> : (2) Multi-level referential integrity and multi level cascading.
> : The ability to link a master table to several sub-levels of child tables
> : is not possible withing Paradox 5.0.  This seems to be an inhibitor
> : to the development of multi-level database structure.
> Can you explain that?  I don't quite understand. An example would be
> helpful...

i provided an exact example for this in the last coupla weeks.  in my
application, there are some tables that should have limited access but
are used as the base references & lookups ie. ISO country codes, valid
department numbers; items that don't change much & should be controlled
by a limited few individuals.  By placing these small tables in another
directory on our Netware server, i am able to restrict access rights
to read & dir list using existing & well organized technology.

however, other tables  - such as inventory or employees - need to be
in a place where everyone as write access and should be using those
same tables in a referential integrity (RI) role - don't want to
record an employee in a non-existant department - to remove the
requirement for OPal programming & to enforce the relationships
*despite* programming attempts.  

Quote:
> : Referential integrity should also be able to be established between
> : tables in other directories and not limited to tables within
> : the master only.
> In Borland terminalogy, Referential integrity is limited within the same
> database, or "directory"...

the result of short-sightedness - whether a deliberate design decision
or not - and resulting in Anthony's (& mine) wish list item for future
development.  If aliases are found everywhere else in the product,
why was it not included in RI?

---------->  text deleted  <--------



Tue, 30 Dec 1997 03:00:00 GMT
 Wish list? or exists?

: (1) Lookup Tables

: It seems that "relocating" a database and/or form within
: Paradox can be an experience in itself.  Hardcoded lookup
: tables (ie drop down edit lists) that make reference to
: Aliases seem to have no difficulty when a form and/or
: database is relocated to another drive/directory.
: However, lookup tables defined in the "restructure" procedure,
: although the alias may be used to define the location
: of the lookup table, Paradox seems to code the PHYSICAL location
: of the lookup table rather than the ALIAS path to the table
: as specified in the IDAPI configuration utility.

This can be blamed on the HISTORY of Paradox, probably. It was expected
that the lookup table be at the same place of the original table...

If you REALLY need lookup across aliases, do it with OPAL.

: (2) Multi-level referential integrity and multi level cascading.

: The ability to link a master table to several sub-levels of child tables
: is not possible withing Paradox 5.0.  This seems to be an inhibitor
: to the development of multi-level database structure.

Can you explain that?  I don't quite understand. An example would be
helpful...

: Referential integrity should also be able to be established between
: tables in other directories and not limited to tables within
: the master only.

In Borland terminalogy, Referential integrity is limited within the same
database, or "directory"...

: A child table, capable of referencing a table which subsequently
: references another master would be an attractive item.

So you want 1:M:1 RI?

--
=========================================================================
Kasey K. S. Chang (a guy)             | Paradox for Windows Programmer

URL> http://userwww.sfsu.edu/~kschang | X-COM Guru   General PC expert
--------------------------------------+----------------------------------
Editor of XCOM and XCOM2 UN-official Strategy Guide -- FREE!  (See URL)



Tue, 30 Dec 1997 03:00:00 GMT
 Wish list? or exists?
[snip]
: the result of short-sightedness - whether a deliberate design decision
: or not - and resulting in Anthony's (& mine) wish list item for future
: development.  If aliases are found everywhere else in the product,
: why was it not included in RI?

This is one of the "shortcomings" I mentioned to Borland when I had a
chance to with members of the Paradox team. Basically, aliases are not
supported universally. Somewhat "uneven".

--
=========================================================================
Kasey K. S. Chang (a guy)             | Paradox for Windows Programmer

URL> http://userwww.sfsu.edu/~kschang | X-COM Guru   General PC expert
--------------------------------------+----------------------------------
Editor of XCOM and XCOM2 UN-official Strategy Guide -- FREE!  (See URL)



Wed, 31 Dec 1997 03:00:00 GMT
 Wish list? or exists?
I would just like to post a thread I recieved from a person
with read only access.

I appreciate all the input, it's nice to see that I'm not the only one
having this problem

CUT______________________________________________

Date: Thu, 13 Jul 95 12:33:56 EST



Subject: Re: Wish list? or exists?

Anthony,

   Boy did you hit the nail on the head!  I've been telling everyone I know
about these two MAJOR ommisions ever since Version 1.0, but I have yet to even
see this addressed by anyone at Borland (i.e. in any of the upgrades).  You'd
think that more people than the two of us would have run into these problems
before, and that they would have been at least mentioned.

Quote:
>  I have a wish list, and since there is nowhere on Borland On-Line to
>  deposit it I think I'll drop it here.

>  If anyone knows of a fix, please let me know, else it will remain,
>  as it was intended to be, a wish list.

>  (1) Lookup Tables...

This one kills me!  There is so much out there about how great aliases are, and
how great table lookups are (both of which I agree), but you can't use aliases
on lookups!?!?  You can use them with everything else!  I haven't even heard of
a workaround for this.

Quote:
>  (2) Multi-level referential integrity and multi level cascading...

Here's what I wanted to do, but couldn't....

          Field Name     Type   Key
          ------------   ----   ---
Table 1:  ProjectName     A      *

Table 2:  TaskName        A      *
          ProjectName     A      *   <---- RI Link to Table 1 (Cascading)

Table 3:  SubTaskName     A      *
          TaskName        A      *   <---- RI Link to Table 2 (Cascading)
          ProjectName     A      *   <---- RI Link to Table 1 (Cascading)

..The last link is the one you cannot do.  It would be nice when a ProjectName
changes, it would cascade the change all the way down to all of the tables.

Quote:
>  Any comments or any more wishes??

How about the ability to re-run Paradox when your original session crashes
WITHOUT HAVING TO EXIT WINDOWS!  I can't tell you how many times, while
developing, I get a GPF and Paradox crashes.  What a pain in the ass it is to
have to exit and re-start Windows.  This _can't_ be that hard to fix.

Maybe Version 6.0 will respond to some of these problems.

Good Luck,
Matt Boge

P.S. I have "Read Only" access to the Internet, so I mailed you my response (let
me tell you how frustrating "read only" is when I'm reading my newsgroups).  You
may post any of this message if you wish.



Thu, 01 Jan 1998 03:00:00 GMT
 Wish list? or exists?
When did you get a chance to meet with members of the Paradox team?
Sounds like fun.

--
tjc


Quote:

>[snip]
>This is one of the "shortcomings" I mentioned to Borland when I had a
>chance to with members of the Paradox team. Basically, aliases are not
>supported universally. Somewhat "uneven".



Thu, 01 Jan 1998 03:00:00 GMT
 Wish list? or exists?
: When did you get a chance to meet with members of the Paradox team?
: Sounds like fun.

Long story, not worth repeating here. Let's just say I took a trip down
to Scotts Valley by invitation, and let's leave it at that.  8-)

--
=========================================================================
Kasey K. S. Chang (a guy)             | Paradox for Windows Programmer

URL> http://userwww.sfsu.edu/~kschang | X-COM Guru   General PC expert
--------------------------------------+----------------------------------
Editor of XCOM and XCOM2 UN-official Strategy Guide -- FREE!  (See URL)



Fri, 02 Jan 1998 03:00:00 GMT
 Wish list? or exists?

Quote:

> Date: Thu, 13 Jul 95 12:33:56 EST


----------------->  text deleted  <------------
Quote:
> >  (2) Multi-level referential integrity and multi level cascading...

> Here's what I wanted to do, but couldn't....

----------------->  text deleted  <------------
Quote:
>           Field Name     Type   Key
>           ------------   ----   ---
> Table 3:  SubTaskName     A      *
>           TaskName        A      *   <---- RI Link to Table 2 (Cascading)
>           ProjectName     A      *   <---- RI Link to Table 1 (Cascading)

> ...The last link is the one you cannot do.  It would be nice when a ProjectName
> changes, it would cascade the change all the way down to all of the tables.

----------------->  text deleted  <------------

actually, i don't fault Paradox for not supporting multi-level
cascaded changes.  in my real life [Pdox is a side-line  ;-] work
with large-scale systems, even IBM's DB2 doesn't support cascaded
changes whatsoever (that a *is* real word, right? :-).  the closest
they come is a cascade option during delete where a parent row
[deletes|nulls|is prohibited by] all the child rows in an RI
hierarchy.  to achieve a cascaded change, you must add the new parent
row, change the child rows, then delete the old parent.  imagine
having to do that over a multi-level RI scheme.  (in truth, there is
an alternate method.  RI can be dropped|added via SQL allowing the
tables to be incorrect for a period of time.)

i can't speak to what Oracle, Sybase, et al allow since i have no
experience outside of DB2 & Paradox.



Fri, 02 Jan 1998 03:00:00 GMT
 Wish list? or exists?

Quote:

>How about the ability to re-run Paradox when your original session crashes
>WITHOUT HAVING TO EXIT WINDOWS!  I can't tell you how many times, while
>developing, I get a GPF and Paradox crashes.  What a pain in the ass it is to
>have to exit and re-start Windows.  This _can't_ be that hard to fix.

For the time being, have a look for the Delphi ?freeware? utility
KILLDLL.EXE (well, actually, it comes as Delphi source, originally on
Compuserve I think). I *haven't* tested it with IDAPI yet (only discovered
it this am), but it lets you list and remove any/all DLLs from your system.
Since IDAPI01.DLL (and poss. some others) are what prevent you from
reloading Paradox, this should do the trick...

I've heard of other DLL killers out there...

cheers,
peter

--
==============================================================================
     Peter Hyde, South Pacific Information Services Ltd, Christchurch, NZ
                    On the whole, I'd rather be flying...



Sat, 03 Jan 1998 03:00:00 GMT
 Wish list? or exists?


[chop]

Quote:
> How about the ability to re-run Paradox when your original session crashes
> WITHOUT HAVING TO EXIT WINDOWS!  I can't tell you how many times, while
> developing, I get a GPF and Paradox crashes.  What a pain in the ass it is to
> have to exit and re-start Windows.  This _can't_ be that hard to fix.
> Maybe Version 6.0 will respond to some of these problems.
> Good Luck,
> Matt Boge

Hi Matt.

When a Windows application has a GPF and closes, *Windows* has
problems with memory management, so it's better to exit and restart.
This is true for every Windows application, not only Paradox.  Maybe
it's only a bit more crashing than others.  :-)  There's no fix to
that, PX should be able to exit clean on crash, which, you'll
understand, is not possible.

I run Paradox on Windows NT, and I don't suffer when it hangs.  It's
NT strength anyway.  Maybe Win95 will ease your life, I can't speak
anyway, I use NT.

     Marco Montagnana

|o|--------------------------------------------------------------|o|
| | Prometeia Calcolo       Phone:      +39-51-237406            | |
|o| Via Marconi 43          Fax:        +39-51-220753            |o|

|o| Italy                   Compu$erve: 100411,2766              |o|
| |--------------------------------------------------------------| |
|o| ... God is real, unless declared as integer.                 |o|
| |--------------------------------------------------------------| |



Sun, 04 Jan 1998 03:00:00 GMT
 
 [ 10 post ] 

 Relevant Pages 

1. F4M I wish I wish.....SMS-solution (Oh Genie help me)

2. SQL 7.0 wish list (tiny)

3. Wish List Email

4. Wish list for version 7.0

5. Something for the Wish List

6. FYI: Informix Wish List

7. DTS Wish List

8. Wish list newsgroup

9. user types API wish list

10. 7.2 wish list (big change)

11. VB 7 Wish List

12. VB7 Wish List


 
Powered by phpBB® Forum Software