In the future . . . .

Discussions on the future of databases, computing , the meaning of life ........

Moderators: Phil Winkler, Graham Smith, Pete Tabord

Re: In the future . . . .

Postby nic » Mon Jan 09, 2012 8:14 am

=D>

Thank you so much for your insights, Pete.

The realization is slowly dawning on me; that my DOS mindset and some of my old Dataease mindset too, is what is making me struggle a bit with coming to grips with Ffenics. I think old dogs like myself, unlearn old tricks with difficulty too. :D

I'm determined to master Ffenics before retirement. (Whatever that means?) :arrow:

Thanks again for a great product.

Kind regards,

Nic
You must never lose your ignorance, you can never get it back again.
nic
 
Posts: 9
Joined: Thu Jul 14, 2011 12:57 pm
Location: Namibia
Has thanked: 0 time
Been thanked: 0 time
 

Re: In the future . . . .

Postby Pete Tabord » Mon Jan 09, 2012 11:04 am

Thanks very much for the positive responses.

I've gone back and tided up the post a little because it was kind of an outburst, but I'm glad you appreciate it.

You wouldn't believe the amount of soul searching I've done over the years about this stuff, because after all I was in charge of delivering DFW (at least up to the last few weeks before release), and its consequences directly or indirectly cost many of my friends and colleagues their jobs, not to mention enormous difficulties for many of our customers.

I hope I've got my tenses right now (learned, learnt...). It's an awful long time since I did grammar!
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby Graham Smith » Mon Jan 09, 2012 3:29 pm

Pete Tabord wrote:Some of the suggestions made at the time were complete nonsense and further discredited the process, to the extent that Sapphire started down the ultimately abortive 'interoperability' route. (Which ends up with an app twice as complex but still dependant on DFD).

They say that confession is good for the soul. This starts in 1991 and my memory of the chronology is a bit hazy, particularly since there were often two or three things going on all at the same time.

As one of the few deeply involved alpha/beta testers, I have to lay claim to being one of the people who pushed for interoperability - hard. Looking back on this now, I can see that it was an ultimately futile and counter-productive position to take, but hindsight is often 20/20. Perhaps more important to what was done is the issue of why it was done. Afterall, you cannot understand past mistakes unless you understand what led up to them.

It would take a dozen pages or more to really tell the story, but suffice it to say that during the development of Express, I made frequent trips to Trumbull to get the latest copy and give them reports on any problems I had found. I had some real issues with the OOP stuff but Arun was convinced that things should be done this way for some reasons that went completely over my head. But what I really had issues with was the lack of DQL. This did not seem to be of any concern for Arun because he saw Express as just a first step and the next version would include a completely new programming language. He was also talking at that time about Express being a new approach to database management as a front end to multiple back ends. It's worth noting that at that time, Microsoft expressed some interested in the program. How much, I don't know, but there were ideas found in DFD and Express that appeared in Access when it came out. In any event, I became enough of an advocate of Express that I went to London for the first UK DataEase Conference and gave a presentation to a full auditorium of DataEase users.

Fast forward to the first tests of what became DFW. IIRC, by this time, Arun had left DataEase and started work on Enterprise Build, which he saw as being closer to his idea of what a DBMS should be. In any event, I think that this is where Pete came in with the mandate of getting DQL into Express. Again, my memory may fail me but I do recall comments being made that the only way this was going to be done without a complete rewrite was with crowbars, hammers, and a lot of duct tape. And despite the best efforts of a lot of people, what we got as a result was the programmatic equivalent of a platypus - something that was not entirely one thing or the other.

It was about this time that “Interoperability” with DFD 5.x became a big issue. This followed naturally from the idea of using DFW as a front end for multiple back ends. But more than that it was seen as a way of transitioning into DFW as opposed to having to do a complete migration. The idea was not without merit because there are still a surprising number of DFD apps still out there that never made the move. Unfortunately it didn’t take too long to discover that this idea was creating a development bottleneck since there were things that could not easily be done without changing the structure of the data tables and you couldn’t do that while you were tied to the old structure. Even more unfortunately, enough people were convinced that interoperability was necessary that crucial progress was stifled. In my defense, while I was one of the people who pushed this early on, I was also one of those who quickly turned against it.

Somewhere is all of this, there was DFD SQL Connect. As concepts go, this was a doozy. Truth be told, the MS/Sybase driver worked pretty darn good once you figured out that you needed to do all the structural work on the back end and then attach DataEase to it. The failure to make this mandatory was, in my opinion, something that significantly hindered acceptance of the product. As proof of this theory, I spent about three years successfully working with a few companies on their DFD/SQL apps. These were large wide area network applications mind you. In any case, I think this was going on at the same time as the interoperability fiasco and the two had a certain synergy. What ultimately doomed both interoperability and SQL Connect was the one inevitable fact that things change. DFW had to change and SQL drivers did change and nothing was able to continue to work together for long.

We’ll skip over ODBC and OLEDB and the new DFW 6.5 DataEase Connect drivers and the ill fated ConnectEase and LinkEase and all the other attempts to play at being Peter Pan and never grow old.

Fast forward to DFW 7, here was the opportunity to revise some of the biggest issues with the program but as had become pattern, rather than fixing things, the drive was to add new and try and force some things to work even if they had already proven to be a bad idea. Consequently, we saw more time wasted trying to make the OLEDB Sapphire Driver work to so that people could put DFW 7 on top of DFW 6.5 rather than just migrating. It was Déjà vu all over again.

When Ffenics came out, we were finally free of any attempts at interoperability but we are still tethered to many old ideas and old ways of doing things. But that’s another story for another day.
Graham Smith
DataSmith, Delaware
"For every expert there is an equal and opposite expert.", Arthur C. Clarke (1917 - 2008)
"X-Clacks-Overhead: GNU Terry Pratchett"
User avatar
Graham Smith
 
Posts: 2501
Joined: Fri Sep 07, 2007 11:31 am
Location: Delaware, USA
Has thanked: 0 time
Been thanked: 1 time
 

Re: In the future . . . .

Postby Pete Tabord » Tue Jan 10, 2012 9:20 am

Hi Graham

As you correctly point out, there were many strands of things going on at the same time. In my previous long posting I have necessarily simplified, because to go in the whole situation would require writing a book.

Arun was a genius with an ability to present his ideas well, but sometimes this made for over-simplification. He believed that Microsoft's influence was such that it would become compulsory to keep all data in an SQL engine, but since he regarded SQL as fundamentally flawed (along with other major players such as Chris Date and Clynch Salley) he wanted PRISM to sit on top of SQL and 'make it work' . This was one of the purposes behind the 'new' database engine that eventually came out in Express. The actual data storage and retrieval is done by code that was (at first) essentially the same as DOS (actually taken from the character based OS/2 version) but there was this fundamentally new intermediate layer called the Multiview, which was intended to not care whether its data came from DOS files or SQL selects.

Various companies were interested in purchasing DataEase, including Microsoft, but I think the stumbling block was that Arun wanted to retain technical control of the product.

Interoperability didn't seem a bad concept, but it totally failed to work out in practice for the intended purpose of gradually moving people over to DFW entirely. There were many other ideas at that time that should have been thought through far more before development time was devoted in these directions instead of improving core functionality. Another minor example of such diversion away from important basic features was the attempt to copy the character based output of DFD in the Assembler-like 'DQL Export' . As a consequence the real DFW export facility is still under-developed now.

Of course Ffenics still is influenced by decisions from the past, both good and bad, but since I decided to retain the Multiview for its major strengths then we get its oddities also. There is no such thing as a perfect database product, so its a question of balance. The primary thing the Multiview gives us is the ability to navigate around the data and drill down in a relatively ad-hoc way without coding. The combination of the Multiview and the basic file-based database engine also means we are highly optimised for the 'Join' function, which allows us to do what really are very complex relational operations with very good perfiormance. This is incidentally why those who try to move a 'native' DFD/DFW/FF app to a SQL back end have performance and design problems. SQL is not very good at complex joins.

As you also point out, and as Arun intended, DataEase/Ffenics as a front end tool on an app whose schema has been developed in SQL works well. There are a couple of Ff apps out ther that do exactly that. It's a pity the other method of developing the app entirely in DataEase over SQL was ever permitted. in any case, that isn't our prime target market for Ff, as that 'front-end' market is now long established and both difficult and expensive to penetrate.
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby Graham Smith » Tue Jan 10, 2012 1:33 pm

Pete Tabord wrote:Another minor example of such diversion away from important basic features was the attempt to copy the character based output of DFD in the Assembler-like 'DQL Export' .

<standing up, raising hand>Guilty. That's another one of the things that I lobbied hard for. I would regret that decision if it were not for the fact that without it, I probably could never have done several of the apps that I did. I still can't do without it. But that's a topic for another day.
Graham Smith
DataSmith, Delaware
"For every expert there is an equal and opposite expert.", Arthur C. Clarke (1917 - 2008)
"X-Clacks-Overhead: GNU Terry Pratchett"
User avatar
Graham Smith
 
Posts: 2501
Joined: Fri Sep 07, 2007 11:31 am
Location: Delaware, USA
Has thanked: 0 time
Been thanked: 1 time
 

Re: In the future . . . .

Postby Pete Tabord » Tue Jan 10, 2012 3:01 pm

Again, it seemed like a good idea at the time. To some people, anyway :wink:

Look, we've done it again. This topic is supposed to be about the future but we end up raking over the past. My fault as well. My long post of a page or two ago was supposed to be about why I try not to do that any more and just get on and build on what we have!
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby lumberjackshaw » Wed Jan 11, 2012 12:06 am

Look, we've done it again. This topic is supposed to be about the future but we end up raking over the past.


How does that saying go..."Those that do not know their history are doomed to repeat it."
:)
--Josh

Associate with men of good quality if you esteem your own reputation; for it is better to be alone than in bad company.
-George Washington
lumberjackshaw
 
Posts: 207
Joined: Mon Jan 17, 2011 8:07 pm
Has thanked: 0 time
Been thanked: 0 time
 

Re: In the future . . . .

Postby Pete Tabord » Thu Jan 12, 2012 9:39 am

While agreeing with the sentiment it is possible to overdo the analysis of the past to the point where complete paralysis results :)

Most of DataEase's past revolves around the percieved problems and solutions of the early 90's.

While I think its obvious to most people entering data that the old CUI systems were very efficient and cost-effective, it's not those people who make decisions. The decision makers want pretty presentation and so on. So the natural result 20 years on is that we have systems that are in many ways less efficient but which please the decision makers. I can't fight that, I can only try to come up with software that tries to make working within that environment as easy as possible. Its no good saying xyz was easier in DOS, or that systems were quicker to write and easier to maintain - if that mattered to the decision makers they'd still be using it.

And the further consideration we now have - have had for at least 10 years really - is that the younger generations have never heard of DOS and regard anything that looks character based as being on a par with stone carving.

In the UK at least - it occurs to me I might have mentioned this before - pretty much everyone leaving school has had exposure to Windows and expects to use Windows keystrokes, menu structures etc. They are very resistant to being retrained to use something which to them looks prehistoric. Windows itself to them is pretty basic and even a tad old fashioned.

And they are right - whether we like or not if we are to survive comemrcially we have to recognise the wishes and prejudices of the environment we operate in - clinging to the past, whatever its merits, simply shuts you out of the conversation. As indeed happened to DataEase in its big business accounts.
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby lumberjackshaw » Thu Jan 12, 2012 12:06 pm

Good Day,

Here is an editor I use for a lot of coding projects (several languages supported). It's a free product that works better than most commercial ones (largely due to simplicity). The reason I have posted this link is because of the editor features it has. It has color highlighting for key words (red for unrecognized, and typos ), automatic indention ( returns to the start of the previous line), statement matching ( placing the cursor on a keyword ,if..end, that has termination highlights both), and quote insertion (if you type a quote it inserts the match after the cursor). I thought it might a good example of ideas but it's worth playing with for your other projects if nothing else. :-)
The one feature it does not have is code completion but it is very fast and works well. The formatting options would (I would think), the simplest and quickest to implement.

http://www.pspad.com/


Enjoy!
--Josh

Associate with men of good quality if you esteem your own reputation; for it is better to be alone than in bad company.
-George Washington
lumberjackshaw
 
Posts: 207
Joined: Mon Jan 17, 2011 8:07 pm
Has thanked: 0 time
Been thanked: 0 time
 

Re: In the future . . . .

Postby Graham Smith » Thu Jan 12, 2012 1:21 pm

lumberjackshaw wrote:Here is an editor I use for a lot of coding projects (several languages supported).

Actually a fellow named Malcom Bolton was working for Sapphire on modifying the editor to behave something like this. Not sure what ever became of that.

There are actually a number of editors that can do things like this. I use one called NotePad++ and you can actually create custom "languages" in it that will highlight "key words". I thought about doing this for Ff but never really found the need.
Graham Smith
DataSmith, Delaware
"For every expert there is an equal and opposite expert.", Arthur C. Clarke (1917 - 2008)
"X-Clacks-Overhead: GNU Terry Pratchett"
User avatar
Graham Smith
 
Posts: 2501
Joined: Fri Sep 07, 2007 11:31 am
Location: Delaware, USA
Has thanked: 0 time
Been thanked: 1 time
 

Re: In the future . . . .

Postby Pete Tabord » Thu Jan 12, 2012 1:44 pm

That was for Net Plus, I believe. And more importantly it was never delivered :(

I thought about doing this for Ff but never really found the need.


There is a twofold advantage to highlighting the different parts of the language in the editor, first the obvious one of having an environment more like, say, visual studio which many people will be used to, but secondly and more important if we can highlight it we can also do more creative things like right click to bring up options such as show me all uses of this thing in the app. (that is, go off and run one of the more info reports but using what you selected as the criteria).
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby Fred Kingston » Thu Jan 12, 2012 2:00 pm

Pete said:
As indeed happened to DataEase in its big business accounts.


Interface isn't what killed DataEase in big business accounts.. :lol: :lol:
Fred Kingston
 
Posts: 281
Joined: Sun Aug 01, 2010 10:54 pm
Has thanked: 0 time
Been thanked: 0 time
 

Re: In the future . . . .

Postby Pete Tabord » Thu Jan 12, 2012 3:31 pm

In so far as it was ever in big business accounts in the US the failure to quickly produce what the head honchos wanted in a Windows product was a major part of it.

Of course what the guys entrusted with doing the buying wanted and what the actual people developing databases wanted were often very different things.

Now there's hardly anyone doing any database development at all in many of those places.
Peter J. Tabord
Head of Development
Database Software Ltd.
ptabord@ffenics.com
Pete Tabord
 
Posts: 1881
Joined: Fri Sep 07, 2007 12:48 pm
Location: Caernarfon, Gwynedd, UK
Has thanked: 0 time
Been thanked: 3 times
 

Re: In the future . . . .

Postby Phil Winkler » Thu Jan 12, 2012 4:50 pm

You are quite right about that, Pete. Many places have actually resorted to spreadsheets and what they call databases in Access. In truth these are simply unnormalized flat tables. We received a "membership database" from a large diving organization. No unique field(s), repeating columns, etc., etc. No wonder they were embezzled twice over the years. No accountability at all.
Phil Winkler
PLM Consulting, Inc.
pwinkler@plmconsulting.com
Phil Winkler
 
Posts: 889
Joined: Fri Sep 07, 2007 12:45 pm
Has thanked: 0 time
Been thanked: 0 time
 

Re: In the future . . . .

Postby Fred Kingston » Thu Jan 12, 2012 6:27 pm

Pratt-Whitney at one time had over 300 unique DataEase databases on their network in West Palm..

Both Pratt and DataEase are gone..
Fred Kingston
 
Posts: 281
Joined: Sun Aug 01, 2010 10:54 pm
Has thanked: 0 time
Been thanked: 0 time
 
 
PreviousNext

Return to Philosophical Discussions

Who is online

Users browsing this forum: No registered users and 1 guest

cron