I know this type of question has been asked numerous times but...

What are the best, most cost effective ways to interface to D3 Windows and
Linux from a web server (Apache)?

We want to host a web service for our D3 systems to provide pricing,
availability or whatever data is needed from D3 to an outside web site. The
outside web site will be doing the presentation and shopping cart stuff. So,
we are only interested in serving data at this time.

I will be running Apache locally (Windows or Linux) to accept incoming
requests, then I need to create a connection to D3 to run some programs to
process the request and return some data.

We are successfully doing this type of thing using JD3
(Apache/Tomcat/Java/JD3/D3) for one small project and it works fine, but I
believe there are scalability issues and it seems like overkill. So, I know
and accept that D3 user licences will be consumed but it has to be minimized
or else it becomes too expensive, and licensing connectivity software (like
Flashconnect) is not going to happen either.

Thanks to all in advance,

Steve

Thanks for the responce Glen.

We are mostly using the web server because we will be getting web requests
from other web servers. So, I will look at Flashconnect.

Thanks,

Steve

"Glen B" <no$pamwebmaster@no$pamforallspec.com> wrote

Steve, I have increasingly less patience for requests where people
want sophisticated functionality and then they end the request with
"but we're not going to pay anything for it". The bottom line is that
you can spend a lot of time getting information and writing your own
tools, or you can pay someone who has made it easy for you. The cost
to buy could be equal to a couple hours of your time where there
entire manual effort itself could go on for weeks. If you want your
business to survive then you need to invest in it with time or money,
and pay for your investment through your profits. I don't mean to
hassle you in particular here, I know something of your history with
D3 and I understand your situation. I'm speaking more in general
terms. Because people are so anxious to jump to that "free beer"
line, I and many others are much less likely to provide any kind of
helpful advice here.

All that aside, to answer your question:

- See this posting I made to the RD forum about using PHP with
FlashCONNECT. The code is there, I have a client who has happily been
building around this without my help for quite a while now.
http://forums.rainingdata.com/index.php?showtopic=869

- At this link you'll find free source code that shows you how to get
into FC via Java. This is the same technique used for PHP above. I
wrote this stuff six years ago and it's still as valid today as then:
http://flashconnect.rainingdata.com/...ava/index.html

- Here is another free interface written by John Lombardo for Perl to
D3, albeit again via FC:
http://tinyurl.com/82zwb

- I've written a client socket interface from D3 which extends to a
middle-tier .NET socket server. Once the connection is made out from
D3, anything can connect into the middle-tier like web services, etc.
This was written as a client rather than as a server because of often
discussed issues with D3 as a socket server. My interface is stable
and benchmarks well, and there is no need to purchase FC, or
third-party .NET controls. The code is in testing at a major client.
With this and using the Visual Studio Web Services wizard you can have
a Web Services Server (WS-S) setup in minutes. I haven't worked out
licensing yet but will be looking for more beta sites soon.

- Check out my series of articles on Web Services at the following
link, there are many links to methods, products, and freeware
solutions:
http://remove-this-partNebula-RnD.com/articles/

- For a completely *nix solution, look into Mono, which works well
with Apache. Here is an article on the topic that may help lead
specifically to free WS solutions - it also has some non-Microsoft
perspective on .NET which may be helpful for those of you who hate
Microsoft and avoid .NET because of the perception that the two are
linked:
http://www.novell.com/coolsolutions/feature/11227.html
About 433,000 other references can be found here:
http://www.google.com/search?hl=en&q=mono+"web+services"
That list includes this helpful page:
http://www.practical-tech.com/middleware/m052302.htm
And this VERY helpful page with a complete tutorial and full source:
http://tinyurl.com/8dkr3
And others related (some .NET, some Mono):
http://www.codeproject.com/cpnet/

- If you want to do this completely for free but with some effort on
your part, here is a fishing pole but no fish:
---- Get or write a standard WS-S using your favorite tools, Perl,
Java, PHP, whatever.
---- Write BASIC programs and a d3tcl command line to satisfy inbound
requests.
---- Get or write an xinetd interface (check mvWWW?) which will serve
as a conduit between the WS-S and your d3tcl command.

This last solution can be implemented using some of the other code
referenced above.

I hope that's a decent start for you. I've written Web Services
clients and servers for our clients and can do the same for you on a
contracting basis. If I use my tools you'll have it fast and at a
decent price. If you insist of specific tools as some of my clients
do, the cost goes up as it takes time to get all of these things to
play nice together.

HTH,
Tony G
TG@ removethisNebula-RnD.com

Tony,

I knew I would catch a flame from my request

I have read through most of your acticles on WS on your web site. It is very
detailed and informative as usual. I will contact you offline for more
information.

Steve

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

Tony Gravagno wrote:
Tony:

The underlying value of FOSS is equally misunderstood.

http://lxer.com/module/newswire/view/53255/index.html

The fact is that getting to specific cost elements is an excercise in
futility. The only way to get extreme productivity is through (shared)
ownership of every piece of the source code and than determine how that
deployment affects the overall productivity of the enterprise.

The fact is that, while Lean Manufacturing has become the norm, overall
productivity has not increased correspondingly because Lean
Manufacturing has come at the cost of Fat IT.

Perhaps I am the only person that posts on these sites who actually
owns a company that uses Pick and I know that I have never seen
anything that comes from Microsoft that does not cost me more than the
return I get. The big problem with Microsoft stuff is that it is like
a switch. It is either on or off. When on the other hand, I deply
Open Source products I can tweak that software in numerous ways to see
how it affacts productivity.

Henry Keultjes
Microdyne Company
Mansfield Ohio UA

I think that there are several of us who have other business interests. I
can't remember a particular instance but I have seen posts that at least
imply such. My family and I have other interests but no manufacturing at
the moment. But your point is well taken - we do better with good old Pick
than we do when we wander off into the world of SQL Server and such. As a
Pick seller and as a Pick user I have rarely actually suffered from - but
often wondered at - the fact that it is the world's best kept secret.
BobJ

<csigline (AT) hotmail (DOT) com> wrote

I am jelous of you Bob because you have the next generation also
working with Pick.

Indeed, Pick is the world's best kep secret but, interestingly enough,
last night at a party on the occasion of Holland sponsoring the
Cleveland Flower Show, I run into this software guy who not only
understood when I mentioned Pick, he actually knew Dick and spoke very
highly of both Pick the man and Pick the system.

Pick has always been unique in it's capability to run everything IT in
a company. Give me enough time to live and I will prove to the world
that Pick can once again inherit the IT world with that same concept.
What bothers me though that there are lots of supposedly Pick people
who are trying to convince other Pick people to give up on Pick.

As to me and my house, that will not even be when I am dead because
someone will inherit the Pick earth from me.

Any takers?

Henry Keultjes
Microdyne Company
Mansfield OH USA




j

csigline (AT) hotmail (DOT) com wrote in news:1139088115.033973.38590
@o13g2000cwo.googlegroups.com:

I don't know if I'd go so far as to say that Pick can "run everything IT
in a company." Pick is great for what it is, but IMO the key is knowing
what solutions are best suited for the problems at hand. As for one-stop
solutions, Microsoft probably has a better handle on that than Pick.

Sure, a lot of us (myself included) depend on Pick to continue to pay the
bills, but it's certainly a good thing to realize Pick's place in today's
technology. As has been said before, it's simply another tool in one's
toolkit.

Regards,
Joe

Joe wrote:

I don't know if I'd go so far as to say that Pick can "run everything
IT
in a company." Pick is great for what it is, but IMO the key is
knowing
what solutions are best suited for the problems at hand. As for
one-stop
solutions, Microsoft probably has a better handle on that than Pick.

Regards,
Joe

Joe:

Obviously you also see value in the one-stop solution. That Pick can
also be such a one-stop solution, "run everything IT in a company",
is very simple to understand. I will limit myself to explaining this
from the standpoint of using PowerPC because that's where my focus is
http://www.ncolug.org/ppc.htm

Ultimately, whatever the application is, it runs on the instruction set
of the processor. So any application that runs on machines that are
PowerPC based like the Apple G5 or the IBM pSeries, iSeries or zSeries
can be made to not just run on Pick but those applications can also be
made to run faster than they typically do now on some kind of OS
because the overhead is so much less.

Obviously running Pick natively on PowerPC requires some real
programming with C and/or C++ and will require additions and/or
modifications to those compilers. Doing so will also require additions
and/or modifications to the DataBasic compiler but as a whole, those
solutions are simpler and more cost effective than the Rube Goldberg
solutions that have evolved to achieve the same objective.

I am 100% convinced that, had there been Open Source solutions like
drivers in the 1986 realm when OA was ported to AIX on the RT, Dick
Pick would have been able to stick with the "everything on Pick"
concept that had made Pick so extremely competitive as a business tool.
It is a real shame that this genius did not live at least another ten
years to take another shot at the IT crown.

Henry Keultjes
Microdyne Company
Mansfield Ohio USA

I have to say I agree totally with Henry here. People seem to forget
native Pick was WAY more than a database.It seems that while everyone
was busy arguing that Pick was at least as good an Operating System as
Unix, relational databases snuck up on the inside and stole the
database crown. Pick ended up as an also ran in both the operating
system and database fields. It could have been so different.

And don't forget Microsoft were talking to Dick Pick before he died and
they went the Access route.

Still, Pick quietly supports plenty of businesses who really care about
efficiency and value for money. The big problem is once those
businesses grow past a certain size, get floated, start having turnover
of line of management staff, and get invaded by graduates, Pick rapidly
becomes a "Legacy" - a term synonymous with "Liability" to those
aforementioned grads.

If only ...

csigline (AT) hotmail (DOT) com wrote: