Hello,

Version 9 of D3 has been out for a few months now... Are there any
pioneers out there that have migrated that would care to share their
experience? Good, bad, or ugly?

I did hunt around for other posts on this subject, but was not able to
find much.

My apologies if this is a repeat. Send me link to something I missed
please.

Thanks

- Jim Laner

Jim, in a recent thread titled "D3 Windows to D3 Linux" I pressed for
anyone to provide testimony regarding specific issues with D3v9 that
prompted a downgrade or migration. Despite a lot of valid anecdotes
about v7 issues, and preferences about OS, no one (IIRC) had anything
bad to say about this most current release. Does "no bad news" mean
everyone has migrated away? LOL Draw your own conclusions.

Personally I use both D3v7 and v9 on Windows and Linux, both for
myself and at client sites. Aside from some enhancements I find
nothing significantly different between 7 and 9. I trust that they've
addressed a number of stability issues and this is what gives many of
us so much less to talk about. I'm not saying this release is
perfect, and I think I've already admitted to forgetting a few issues
that I've found, but these things come and go with all platforms. I
feel as comfortable with the stability of D3 as I do with any other
system - and I use almost all of the MV variants out there.

When I think about D3 Windows and the FSI compared to VME
environments, I do wonder about how scalable indexes are and exactly
what we can or cannot index. There are still many verbs that are
different between platforms and subtleties about how things work with
Flash. As much as I'd love to use "jBase-style" D3Flash EXEcutables
(see "d3flash" and .D3f files in docs) I don't because I'll be
frustrated by its absence in *nix, and with limitations in Windows.
But I don't see these things as reasons to migrate away. And for the
record, in my experience, most (all?) of the other MV platforms have
exactly the same kind of differences across OS's.

I'll take that further and speak up here where others may not: I think
a number of people have migrated from D3 and found exactly the same
cross-OS nuances in other platforms. But they put so much time and
money and reputation on the line for their big change that it would be
way too embarrassing to state publicly that in many ways their
migration was simply a side step and not a real step forward. There
are also people who have gone from D3/Windows to something like
QM/Linux and they are extremely happy. The same happiness might have
been attainable in a simple D3 OS change - which is exactly what a lot
of people here say they've done with great success. YMMV

I encourage you to take your inquiry to TL Support managers - or meet
with some senior management who are now closer by you. I know these
people well and have never had reason to question their responses to
inquiries about open issues. While they rightfully will/should never
tell you if or when some issues are going to be addressed, they will
acknowledge any issues that have been reported and advise you
regarding how such issues may relate to your environments.

HTH,
T


Jim wrote:

Hi Jim.

I am a long time user of D3 in Australia. Due to the unreliability of
D3 on Windows during 7.4 , I pulled all my customers either to D3
Linux or Universe Windows.

However, I recently had a need to knock up a quick web based system
for a customer (who was one of the ones ported onto Universe
Windows).The customers system is a Windows 2008 SBS with exchange.
A trial of Mysql & ASP failed miserably, so I rewrote the system in
Flashconnect 3.8 on D3v9. Tests on my development system and
subsequently after 12 months of live deployment have shown the system
to be rock solid. All of the files reside in the VME (with some OS
level files shared with Universe). It has extensive indexing and uses
triggers to replicate files to Universe. Touch wood, I have not had to
do a VME restore since installation, and it runs happily without any
user intervention. Only problem I had during installation was with
Java versions (it uses a different version to Universe, but the local
distributor T-Data helped to overcome the issues). Being a
flashconnect system, all the programs are flash compiled.

If I was recommending a new system, I would still go Linux, but for
existing Windows users D3v9 is back on my list of supported databases.

HTH
Regards,
Noel

On Aug 20, 8:28*am, Jim <jimb... (AT) yahoo (DOT) com> wrote:

Noel, I am curious why Mysql & ASP failed miserably. It does seem to be
quite a popular platform.
Thanks,
Jeff

And was that decade old ASP, or current ASP.NET?

"Jeff Caspari" wrote:

I am reasonably experienced in MYSQL in Linux but not very experienced
in ASP.net. When the customer asked for the system, I had a
development tool called IronSpeed that I had used for another project,
so I used it to generate the screens and reports. When the resulting
system was rolled out, it was so slow as to be unusable. Also the Ajax
callback was very unreliable (the screens needed to do pricing updates
during line item entry). Over a weekend I rewrote the system in
Flashconnect (using D3 Linux), showed the customer that it worked then
ported it to D3 Windows and never looked back.

I have other projects that I have written using VB.net & mysql and
don't have any speed issues (or any other issues for that matter). It
may be that IronSpeed did not generate very efficient code for MYSQL,
perhaps with MSSQL it may have been better.

Noel

On Aug 22, 4:56*am, Tony Gravagno <tony_grava... (AT) nospam (DOT) invalid>
wrote:

*smile* OK, I won't point fingers here but this is an interesting
situation.

Google for "iron speed slow data" and you'll get a lot of hits.

You used a tool to avoid code. Performance was sacrificed for shorter
development time. We can hope that these miracle tools that do
everything for us will do so efficiently, but they often generate
inefficient boilerplate code for all possible scenarios. It's like
fast food which usually doesn't taste as good as homemade.

When you used tools with which you were familiar, and where you
actually wrote the code, you had better success. So with more
investment you had better success in deployment. The problem wasn't
MySQL, nor ASP, nor ASP.NET. As much as I'd like to hop on the D3
back-patting parade, D3 (and v9) can't be lauded with this anecdote.

I'm focusing on this for a couple reasons - again, not to point any
fingers, Noel. Some people use tools to avoid learning the platform
and then they blame the platform. Pick people who have never become
proficient with another coding language often talk about how bad
everything is outside of BASIC. Someone could blame D3 if they didn't
like FlashCONNECT. They could blame Universe if they had an issue
with UniObjects for Java (UOJ).

With DesignBais, people must accept that they have no control over how
things work when they decide they don't want to know how things work.
That's not a bad decision, it's just part of the price. The same
could be said to an extent for mv.NET, for which in full disclosure,
I'm a vendor.

Indeed, many people have found ASP.NET to be slow - it's a tool that
separates us from "the metal". Many people blame Windows and
Microsoft - the platform rather than the tools. And some jump to
PHP/Zen or Java/Tomcat and then praise Linux ... go figure. Many
ASP.NET devotees cite WebForms as the problem because that environment
does so much for us. Many have switched to MVC to get better
performance through more control. Personally I haven't made that jump
yet - I've learned to optimize a lot of ASP.NET code when I have the
source to do so, and I've chosen to let WebForms do more for me,
sacrificing a little performance so I can avoid more rigorous coding
with MVC. Again, knowledge helps us to make better decisions.

The real solution is to take the time to learn how things work so that
we can make better decisions about the tradeoffs we'll accept when we
use tools to save us time.

Or as Admiral James T Kirk said:
"You have to learn Why things work on a starship."

Tony Gravagno
Nebula Research and Development
TG@ remove.pleaseNebula-RnD.com
Nebula R&D sells mv.NET and other Pick/MultiValue products
worldwide, and provides related development services
remove.pleaseNebula-RnD.com/blog
Visit PickWiki.com! Contribute!
http://Twitter.com/TonyGravagno




Noel Borel wrote:

Just to clarify some points. Before this project, I had used Ironspeed
for 2 other projects. In both cases it worked without a problem
(although both used MSSQL), but they were in-house applications
running on an internal network. For this project the Ironspeed
application was being access from outside the organisation. Amongst
other things, it did not seem to handle latency very well.
Flashconnect handles it without a problem.

Tony, in recent discussions you have talked at length about Pick
programmers getting out into the real world. So your comment "When
you used tools with which you were familiar, and where you actually
wrote the code, you had better success. So with more investment you
had better success in deployment. The problem wasn't MySQL, nor ASP,
nor ASP.NET. As much as I'd like to hop on the D3 back-patting
parade, D3 (and v9) can't be lauded with this anecdote. " suggests to
me that I have not I should not have used this tool until I had made
more of an investment in knowing how it worked. Based on that
criteria, I would never have used Flashconnect 10 years ago, as I
only had 3 days training with Mark Brown in Pick Systems offices
before coming home and writing a system. By the way that system just
had its 10th anniversary last week.

My original anecdote was about the reliability of D3v9. Personally, I
feel D3v9 can be back patted and lauded, but everyone is entitled to
their own opinion. The accompaning comments I made have been used to
generate a soapbox platform for a purpose I don't understand. I will
now return to the shadows.........

Noel - You made two separate statements:
1) A trial of Mysql & ASP failed miserably
2) D3v9 is back on my list of supported databases

This subthread has nothing to do with D3.
Jeff said about your point #1:
"I am curious why Mysql & ASP failed miserably"

You answered that you used a tool that failed you. I commented that
sometimes people fault a platform when tools fail, and sometimes when
one platform is faulted, it may appear that consequently another
platform is better. I'm trying to separate a perception of cause and
effect there.

I like D3 as much as you do. But I also wanted to be fair and point
out that while tools failed you for whatever reason, it doesn't make
them bad tools individually or as a whole. IMO, that is the
perception that was left by your note, and I believe it prompted
Jeff's inquiry.

Nor does failure of those tools give anyone who might be reading here
justification to tout MV or D3 as superior, citing your experience as
an anecdote. It's easy to see someone doing that, and in this case I
don't think that would be fair or reasonable. A D3 VAR will lose a
sale if he boasts MV superiority while citing the miserable failure of
other tools. Regardless of the competence of D3 or the application,
someone who knows those tools will feel insulted that this tactic is
being used and it will derail the sales process.

THAT is my ultimate point.

Your success with D3 and FlashCONNECT stands on its own. Kudos to
you, and yes, "D3v9 can be back patted and lauded". But I think we
should be more reserved about raising such fanfare on the backs of
other platforms. If we use this approach, we'll lose in "real"
discussions.

I hope that's helped to clarify.
T

The reason I asked was because we have been doing a lot of development using
asp.net and sql and have experienced incredible stability. I was wondering
if there was any inherent problems we haven't seen. I'm glad it was mainly
a problem with the development tool.