Re: Another multivalue synonym question

#1

I have to smile. I maintain to this day that this issue is probably
the absolute most requested feature for D3 and it's never been
implemented purely for historical reasons. Some people ask why the MV
model doesn't evolve. It's because of things like this - that the
virtual modes in there are so old, obfuscated, and fragile that no one
wants to get in there and take a chance on messing them up. It's fear
that limits progress, not lack of demand.

As John mentioned, Unidata does return a MV set from a translate.
Then again, that code wasn't written using the R83/AP code base and it
was obvious to those guys that this is a logical feature to build in
from the start. It's a shame PS/RD haven't seen that after so many
years, but no surprise either.

T

#2

Tony Gravagno wrote:

Quote:

I have to smile. I maintain to this day that this issue is probably
the absolute most requested feature for D3 and it's never been
implemented purely for historical reasons. Some people ask why the MV
model doesn't evolve. It's because of things like this - that the
virtual modes in there are so old, obfuscated, and fragile that no one
wants to get in there and take a chance on messing them up. It's fear
that limits progress, not lack of demand.

"PVA considered harmful"? <g>

Quote:

As John mentioned, Unidata does return a MV set from a translate.
Then again, that code wasn't written using the R83/AP code base and it
was obvious to those guys that this is a logical feature to build in
from the start. It's a shame PS/RD haven't seen that after so many
years, but no surprise either.

What does UDT do if the key to the translate is multivalued? Or worse
yet, subvalued?

Luke

#3

Luke Webber wrote:

Quote:

What does UDT do if the key to the translate is multivalued?
Returns subvalues, very nicely.

Quote:

Or worse yet, subvalued?

I do not know (nor care to find out!)
"SVMs considered harmful."

--
frosty

#4

It may be my mind playing tricks on me, but I seem to recall many many years
ago a short-lived release had the translation working such that the value
marks were being returned when a translate was being used.

I so no good reason why RD can't add another letter into the translate code
that would bring back the requested attribute with value and sub-value marks
intact.

RD must have a dozen requests stating such a thing. I guess they take such
request seriously to the nearest shredder?!?!?

Can't say I didn't try, but I have given up.

Regards, Dale

"Tony Gravagno" <g6q3x9lu53001 (AT) sneakemail (DOT) com.invalid> wrote

Quote:

I have to smile. I maintain to this day that this issue is probably
the absolute most requested feature for D3 and it's never been
implemented purely for historical reasons. Some people ask why the MV
model doesn't evolve. It's because of things like this - that the
virtual modes in there are so old, obfuscated, and fragile that no one
wants to get in there and take a chance on messing them up. It's fear
that limits progress, not lack of demand.

As John mentioned, Unidata does return a MV set from a translate.
Then again, that code wasn't written using the R83/AP code base and it
was obvious to those guys that this is a logical feature to build in
from the start. It's a shame PS/RD haven't seen that after so many
years, but no surprise either.

T

#5

Reality and Reality X always returned spaces, I am not sure what the
latest version of reality X does. I also can not remember what Prime
did, tho i suspect it was I types anyway so it would have been fine.

What the translate returns on unidata for a subvalued list i do not
know, maybe @TM delimited ? - as said above i would not normally use
English/TCL on subvalues, but they aint evil, i use them a lot so long
as it is only code that uses them...

#6

The mantra that the PVA code is "obsfuscated" "fragile" etc etc is just a
lame excuse. And the historical reasons excuse is just
another...ah...excuse. For one thing, don't they have PVA coders? For
another, isn't it a bit weird (I was going to say shameful) that a company
would sell a product that it can admit is so bad and unmaintenable (assuming
that it is, indeed, RD that says this).

I can think off hand of a half-dozen people (including myself) who could
implement this feature in about a day or two. So how about it, RD? Can you
ship me a PO and an IP address? You can have it by next Tuesday. TG, please
forward this to the appropriate authorities.

Chandru Murthi

"Dale" <benedictknowspam (AT) silk (DOT) net> wrote

Quote:

It may be my mind playing tricks on me, but I seem to recall many many
years ago a short-lived release had the translation working such that the
value marks were being returned when a translate was being used.

I so no good reason why RD can't add another letter into the translate
code that would bring back the requested attribute with value and
sub-value marks intact.

RD must have a dozen requests stating such a thing. I guess they take
such request seriously to the nearest shredder?!?!?

Can't say I didn't try, but I have given up.

Regards, Dale

"Tony Gravagno" <g6q3x9lu53001 (AT) sneakemail (DOT) com.invalid> wrote in message
news

dsed2hqlpnim1jgll8uq6mnr5gfs8rlc2 (AT) 4ax (DOT) com...
I have to smile. I maintain to this day that this issue is probably
the absolute most requested feature for D3 and it's never been
implemented purely for historical reasons. Some people ask why the MV
model doesn't evolve. It's because of things like this - that the
virtual modes in there are so old, obfuscated, and fragile that no one
wants to get in there and take a chance on messing them up. It's fear
that limits progress, not lack of demand.

As John mentioned, Unidata does return a MV set from a translate.
Then again, that code wasn't written using the R83/AP code base and it
was obvious to those guys that this is a logical feature to build in
from the start. It's a shame PS/RD haven't seen that after so many
years, but no surprise either.

T

#7

To add to the list of products that do this "right"...

QM works much the same way as UniVerse and Unidata. The TRANS()
function can take a mutlivalued id item and will return the results
with the marks demoted by one step so that you can sort out whether the
multivalues are because the id list was mutlivalued, because the remote
data was multivalued, or a mix of both.

Of course, we had to maintain the way in which the T conversion code is
defined to work with this nasty replacement of marks by spaces
otherwise users migrating from D3 (etc) moan.

We added a new function, RTRANS(), that does not demote the marks and
gives compatibility with Reality (or so we are told). Also, rather
nicely, our version of TRANS() can retrieve the value of an item that
is itself a calculated item (I-type or A/S with correlative). Wow, that
makes life easy!!

Martin Phillips, Ladybridge Systems

#8

"murthi" wrote:

Quote:

TG, please forward this to the appropriate authorities.

I can summarize my sentiments as : this aint my business.

Initiate discussions through sales (AT) rainingdata (DOT) com.

T

#9

Martin

Is it possible that you could update your newsreader settings to have your
display name as "Martin Phillips" rather than "None" ?

Thanks
Simon

--
Simon Verona

"None" <MartinPhillips (AT) ladybridge (DOT) com> wrote

Quote:

To add to the list of products that do this "right"...

QM works much the same way as UniVerse and Unidata. The TRANS()
function can take a mutlivalued id item and will return the results
with the marks demoted by one step so that you can sort out whether the
multivalues are because the id list was mutlivalued, because the remote
data was multivalued, or a mix of both.

Of course, we had to maintain the way in which the T conversion code is
defined to work with this nasty replacement of marks by spaces
otherwise users migrating from D3 (etc) moan.

We added a new function, RTRANS(), that does not demote the marks and
gives compatibility with Reality (or so we are told). Also, rather
nicely, our version of TRANS() can retrieve the value of an item that
is itself a calculated item (I-type or A/S with correlative). Wow, that
makes life easy!!

Martin Phillips, Ladybridge Systems

#10

Oh, sorry, I thought you were the cdp spokesperson for RD on these matters?
Must've misread something.

C

"Tony Gravagno" <g6q3x9lu53001 (AT) sneakemail (DOT) com.invalid> wrote

Quote:

"murthi" wrote:
TG, please forward this to the appropriate authorities.

I can summarize my sentiments as : this aint my business.

Initiate discussions through sales (AT) rainingdata (DOT) com.

T