Should we be using NewEra
Alan M Cook
I would like to bemoan the pain we had experienced with NewEra and to get
some advice from you experts out there whether it is worthwhile to continue
with it (and Informix) in the light of the uncertainty that seems to hangs
over the future of NewEra.
Several years ago we used Informix 4GL and had reasonable success with it.
The database was stable, fast and we liked it. Then we had to change to a
GUI environment. We are operating in a scientific environment and with the
advent of GIS, it was required that all new applications must be graphical.
We must develop a number of very important mission critical applications,
which will happen over a number of years. After several products were
investigated, we decided on NewEra (Version 1). We realised that the
product was, at that stage, still immature, but supported OO development,
which put it in a class of its own in comparison with what other DB vendors
could offer. We also had a look at third party front-end tools, but the
combination was expensive and we experienced incompatibilities when it was
tested.
We experienced lots pain to implement NewEra. The first problem was
support. Apart from the relative elementary NewEra training, we had to sort
out most of our own problems. With Version 2 we stepped into the trap of
the Application Framework Class Library. The unsatisfactory interface and
bugs wasted much time as we tried to devise workarounds. Most experts with
whom I had discussions lately, strongly advise users not to use it and
rather to develop your on base classes. The lack of basic easy-to-use
tools, such as tabwindows, was also a time waster - we had to develop or
find such aids. After the struggle with the long and steep learning curve,
little support and building our own base classes, it seemed as if our
problems were easing off. The release of the much-improved Version 3.0 and
now Version 3.1 with OCX support, also proofed that the product has reached
maturity. No such luck! The news of the possible end of NewEra, really put
the provable cat amongst the pigeons. The biggest Achilles heel is the lack
of NewEra programmers and the long time it takes to train them. We are
dependent on contractors and they are now extremely reluctant to invest more
time in mastering a dying product.
We had many discussions about possible alternatives, but came to the same
conclusion as Nils Myklebust. Here is an extract from his reaction:
A significant problem for Jeff Hudson and Informix is that there aren't
really that many development tools out there that can be used for major
(Informix?) database developments. He mentions Visual Basic. Well it's nice
for a small client only applications. When it comes to major applications
that also need server side programming it's not a viable solution unless you
decide up front that Windows NT shall forever be your only server option.
He mentions Java. I happen to believe that Java is and will be a tool for
major developments. There are now several large, to very large, systems
being developed in Java, and it all looks good. There is however still
troubled water ahead, and it will still be some time before it's tested
enough for most purposes a database development tool is needed. He mentions
no other. The reason is simply there are no other major tools for the great
public. Yes there is Delphi, but essentially with the same limitations as
Visual Basic (although it's clearly a better language). PowerBuilder has
the same problems. Then there is Forte. Too expensive, coming from a small
company that feels the competition from Java. Then there is a plethora of
tools from companies all over the world. All are relatively small, however,
generally very expensive, and not an option for many. NewEra was actually
always cheaper than most of these, and now with better functionality than
most. Last there is C or C++. In the database development arena this is a
loosing proposition. If you develop an OS, a database engine or even a word
processor or similar tools there is hardly any way around C. It simply take
s too long to create finished database type applications with these alone.
Many excellent database programmers would hardly be able to start at all.
Now here is the essence of my dilemma:
We have to develop a mayor application, which will take several years, and
we need a reliable supported development tool which will be maintained and
enhanced for a number of years. We are not in the research business. Our
vendor must safe guard our investment by providing transition paths for new
emerging technologies. Peace of mind is very essential part of the deal.
For us using the best database without a proper toolkit is like eating
gourmet soup without a spoon – not enjoyable!
I am afraid that we might have reach the end of a road. Am I overreacting,
or is it time to change vendors before investing more money in a mayor
project which can potentially affect the life of many people?
David W Chan
"Alan M Cook" <alan...@global.co.za> wrote:
>We have to develop a mayor application, which will take several years, and
>we need a reliable supported development tool which will be maintained and
>enhanced for a number of years. We are not in the research business. Our
>vendor must safe guard our investment by providing transition paths for new
>emerging technologies. Peace of mind is very essential part of the deal.
>For us using the best database without a proper toolkit is like eating
>gourmet soup without a spoon – not enjoyable!
>I am afraid that we might have reach the end of a road. Am I overreacting,
>or is it time to change vendors before investing more money in a mayor
>project which can potentially affect the life of many people?
Despite what Informix say, I think NewEra is not the right product to
use. If you want to stick to Informix products then have a look at
Data Director for Java which is effectively NewEra for Java. It uses
the same Window Painter and editor (Codewright) as NE v3. Even the
Visual Class Library is similar. And it costs a lot less than NewEra.
But if you go down the Java route, you should investigate all the
major development environments. I think that Java is mature enough to
use now. Have a look at www.eaglestardirect.co.uk. This is a Java
applet which does online motor insurance quotes and allows you to buy
it online as well.
We have built a couple of large NewEra systems but we now use Forte
and Java. Forte is expensive but the distributed object aspects of
the product are way ahead of the others and if you need that
capability then Forte is cheaper than writing it yourself. If Java or
Forte are not what you want then you should probably use Delphi.
David
David W Chan
GO5 Limited
da...@go5.com (work)
d...@cix.compulink.co.uk (home)
Mike Segel
David W Chan wrote:
> =
> "Alan M Cook" <alan...@global.co.za> wrote:
> =
> >We have to develop a mayor application, which will take several years,=
and
> >we need a reliable supported development tool which will be maintained=
and
> >enhanced for a number of years. We are not in the research business. =
Our
> >vendor must safe guard our investment by providing transition paths fo=
r new
> >emerging technologies. Peace of mind is very essential part of the de=
al.
> >For us using the best database without a proper toolkit is like eating=
> >gourmet soup without a spoon =96 not enjoyable!
> =
> >I am afraid that we might have reach the end of a road. Am I overreac=
ting,
> >or is it time to change vendors before investing more money in a mayor=
> >project which can potentially affect the life of many people?
> =
> Despite what Informix say, I think NewEra is not the right product to
> use. If you want to stick to Informix products then have a look at
> Data Director for Java which is effectively NewEra for Java. It uses
> the same Window Painter and editor (Codewright) as NE v3. Even the
> Visual Class Library is similar. And it costs a lot less than NewEra.
> =
Hmmm, I would tend to disagree.
You already have a sunk cost in to N.E. =
(If I remember the original post.)
While I am not a fan of N.E, its mainly due to the braindead class =
libraries, not the language per se.
The problem is that there is already a pool of expertise in N.E =
which would have to relearn OO technologies when going to Java.
Plus, there are some things that Java is too immature for.
(Read: Java is not a proven technology.)
=46rom the poster's comment, I don't think that they will be happy
with any tools currently on the market. Perhaps I am reading too much
in to the comments, but it appears that there are other issues
which are driving the decision to question N.E. I would hazzard
a guess that there are management issues which are not being stated.
First and formost, one must get over the idea that there is this
magical OO tool which can be bought for $5.00 and do everything
including slice bread. This is the issue of cost. I mean, in the
real world, UNIX, vi is free. (It comes standard with UNIX.) =
Do you know how many times I have heard developers slamming =
vi? But hey, you get what you pay for, and it does its job
well. (Beats the hell out of ed or cat ;-) My point is that
there are low cost options in tools, however, don't bitch when
they don't do everything and their limitations may have serious
drawbacks!
> But if you go down the Java route, you should investigate all the
> major development environments. I think that Java is mature enough to
> use now. Have a look at www.eaglestardirect.co.uk. This is a Java
> applet which does online motor insurance quotes and allows you to buy
> it online as well.
> =
There are a lot of Java stuff around. The problem is that most tools and
programmers still view Java in terms of applets. There is much more to
Java and its potential. However, Java still lacks in certain areas which
are critical in speed and distributed objects.
> We have built a couple of large NewEra systems but we now use Forte
> and Java. Forte is expensive but the distributed object aspects of
> the product are way ahead of the others and if you need that
> capability then Forte is cheaper than writing it yourself. If Java or
> Forte are not what you want then you should probably use Delphi.
> =
Well, maybe. Truth be told, the most mature, stable, OO product is =
Openstep. We are talking 12 years of OO development. Strong object
hierarchy,
and does a lot of things right. While cost may be an issue, in the
larger =
scheme of things, its a minor point. As far as being way ahead on
distributed
objects, they wrote the book. :-)
Again all the tool recommendations you have mention, lock you in to a
Microsloth
solution, and do not allow for fast reliable server side or distributed
objects.
Sorry to get on my soapbox, but I just spent 2.5 hours with a group of
people who =
were looking at OO tools and development. I had to agree with their
choice of =
Openstep. Its Mach/NT/Solaris/HP-UX based now, and Corba Complient. Also =
did I mention that its a key feature of Rhapsody? Now you are talking =
platform independence unlike a Redmond. WA. Company would want you to
believe
their product is! ;-)
Please also keep in mind that the Financial Industry has embraced
Openstep/Web Objects.
Why? Because it does the job, does it well, and the costs can be
justified.
(If only Informix would have the same attitude towards doing a Linux
port :-)
Just a few rants and raves from your favorite, albeit excentric Uncle! =
-Mikey
-- =
#include <std_disclaimer.h> /* Mike Segel (MS385) */
#include <No_Spam.h>
#ifdef OFFENDED_BY_CONTENT
The author takes no responsibility for this post. =
Any resembalence to a coherent rational thought is purely coincidence.
-The Management.
#endif
*****************************
Due to AGIS's Refusal to Act Responsibly =
We are blocking all of their domains at the packet level.
This block will exist until AGIS modifies their policies to =
conform to existing RFCs and net community standards.
We encourage all ISPs and domain holders to do the same.
*****************************
Dragi Raos
Mike Segel wrote:> > >We have to develop a mayor application, which will take several years, and
> > >we need a reliable supported development tool which will be maintained and
> > >enhanced for a number of years. We are not in the research business. Our
> > >vendor must safe guard our investment by providing transition paths for new
> > >emerging technologies. Peace of mind is very essential part of the deal.
> > >For us using the best database without a proper toolkit is like eating
> >
> > >I am afraid that we might have reach the end of a road. Am I overreacting,
> > >or is it time to change vendors before investing more money in a mayor
> >
> > use. If you want to stick to Informix products then have a look at
> > Data Director for Java which is effectively NewEra for Java. It uses
> > the same Window Painter and editor (Codewright) as NE v3. Even the
> > Visual Class Library is similar. And it costs a lot less than NewEra.
> >
> You already have a sunk cost in to N.E.
> While I am not a fan of N.E, its mainly due to the braindead class
>
> The problem is that there is already a pool of expertise in N.E
> Plus, there are some things that Java is too immature for.
> (Read: Java is not a proven technology.)
>
> with any tools currently on the market. Perhaps I am reading too much
> in to the comments, but it appears that there are other issues
> which are driving the decision to question N.E. I would hazzard
> a guess that there are management issues which are not being stated.
>
> First and formost, one must get over the idea that there is this
> magical OO tool which can be bought for $5.00 and do everything
> including slice bread. This is the issue of cost. I mean, in the
> real world, UNIX, vi is free. (It comes standard with UNIX.)
> well. (Beats the hell out of ed or cat ;-) My point is that
> there are low cost options in tools, however, don't bitch when
> they don't do everything and their limitations may have serious
> drawbacks!
>
> > But if you go down the Java route, you should investigate all the
> > major development environments. I think that Java is mature enough to
> > use now. Have a look at www.eaglestardirect.co.uk. This is a Java
> > applet which does online motor insurance quotes and allows you to buy
> > it online as well.
> >
> programmers still view Java in terms of applets. There is much more to
> Java and its potential. However, Java still lacks in certain areas which
> are critical in speed and distributed objects.
>
> > We have built a couple of large NewEra systems but we now use Forte
> > and Java. Forte is expensive but the distributed object aspects of
> > the product are way ahead of the others and if you need that
> > capability then Forte is cheaper than writing it yourself. If Java or
> > Forte are not what you want then you should probably use Delphi.
> >
> hierarchy,
> and does a lot of things right. While cost may be an issue, in the
> larger
> distributed
> objects, they wrote the book. :-)
>
> Again all the tool recommendations you have mention, lock you in to a
> Microsloth
> solution, and do not allow for fast reliable server side or distributed
> objects.
> Sorry to get on my soapbox, but I just spent 2.5 hours with a group of
> people who
> choice of
> believe
> their product is! ;-)
>
> Please also keep in mind that the Financial Industry has embraced
> Openstep/Web Objects.
> Why? Because it does the job, does it well, and the costs can be
> justified.
> (If only Informix would have the same attitude towards doing a Linux
> port :-)
>
> Just a few rants and raves from your favorite, albeit excentric Uncle!
> --
> #include <std_disclaimer.h> /* Mike Segel (MS385) */
> #include <No_Spam.h>
> #ifdef OFFENDED_BY_CONTENT
> The author takes no responsibility for this post.
> -The Management.
> #endif
> *****************************
> Due to AGIS's Refusal to Act Responsibly
> This block will exist until AGIS modifies their policies to
>
> We encourage all ISPs and domain holders to do the same.
> *****************************
Might I suggest another great product: Prolifics (see
http://www.prolifics.com). The company is small, but they are around
some 12 years, and keep getting better. The product produces 2 or 3 tier
applications, integrates BEA TUXEDO middleware extremely well (see also
http://www.beasys.com), deploys without changes or even recompiles on
character terminals, X, MSW, Mac, and Web (honestly), uses visual
development of both client and server components, and, using something
they call Transactional Object Model deliver on (almost) codeless
development promise (but still include rather powerfull 4GL and C
intergates easily).
OK, enough raves, or you will guess that I distribute the product (but
not on your market, and we use it internally for long time - we realy
are enthusiastic about it).
Regards,
Dragi "Bonzi" Raos
+--------------------------------------------------------+
| 4-MATE Information Engineering Inc. |
+--------------------------+-----------------------------+
| 4-MATE | Tel: +385 (1) 242-116 |
| Kosa bb | +385 (1) 242-126 |
| 10000 Zagreb | Fax: +385 (1) 242-121 |
| Hrvatska (Croatia) | e-mail: bo...@4mate.hr |
| | URL: http://www.4mate.hr |
+--------------------------+-----------------------------+