Well, the larger apps often are. Still. A lot of Kylix conversions also.

It seems that I'm atypical, since even on the job I don't use Delphi for db related purposes. (this small CRM thingy excluded)

For me this analogy goes a step too far. Yes, there is a superficial likeliness (tool vendor bought by db oriented company), but there are differences too.

Like clarion sold 4th generation DB languages, not DB admin and analysis tools. The difference is that clarion products would have competed with Topspeed products (if used for db access apps, and there were dbf toolboxes for it) while this conflict doesn't exist for Embarcadero<->Codegear.

The objective of Embarcadero seems two fold. Earning money, and some synergy due to the bundling of retail channels.

But most importantly is not what things are named or classified, but what really happens:

Clarion stripped the devel team from the compiler teams for own purposes, while Embarcadero actually raised the devel count of Delphi/BCB to a higher number than in Borland times.

And this shows, D2009 and D2010 are better and wider improved than the versions before it (or anything till D6, D7 being more a maitenance release). Except for the help.

The Delphi decline is IMHO threefold: (in this order)
- Microsoft and Sun's multibillion development tools effots (with Codegear being more a tens of miilions magnitude company)
- Has gradually losts its foothold in education over the years. Partially due to natural causes, partially because of neglect.
- the longtime neglect by Borland, due to financial mismanagement (siphoning profits to the ALM stuff) and failed .NET adventures.

Marco van de Voort wrote

It seems that I'm atypical, since even on the job I don't use Delphi for db related purposes. (this small CRM thingy excluded)

I've got quite a lot of database-related stuff, but far more interesting are the areas where a database (e.g. Sqlite or SleepyCat) is embedded in an application where you wouldn't normally expect it (e.g SVN).

In the case of JPI Clarion had a critical interest in the tools since they not only used the compilers but- if I am remembering things correctly- embedded the compiler backend in their own tools. I'm assuming (perhaps wrongly) that Embarcadero has similar reliance, and that both companies concluded that it would be a disaster if somebody else grabbed the technology. It's worth noting that a few other companies relied heavily on the TopSpeed tools, for example Psion (early PDAs) for whom (again IIRC) there were some special hacks in the linker.

Clarion stripped the devel team from the compiler teams for own purposes,

Yes, I knew some of the people involved. There were "goings on" that I won't discuss.

The Delphi decline is IMHO threefold: (in this order) - Microsoft and Sun's multibillion development tools effots (with Codegear being more a tens of miilions magnitude company)

Interesting to add IBM/Eclipse to this. I think that Lazarus competes with Java at least as much as it competes with Delphi.

- Has gradually losts its foothold in education over the years. Partially due to natural causes, partially because of neglect.

Article in the NYT this morning where various people are commenting that MS has similarly lost its foothold- all the interesting startups are based on "free" technology.

- the longtime neglect by Borland, due to financial mismanagement (siphoning profits to the ALM stuff) and failed .NET adventures.

ALM was doomed from the start, but with .NET Borland was "damned if they did and damned if they didn't": if fewer compiler vendors had signed up for it then MS might have felt under more pressure to make it succeed.

Don't have those either. We are thinking about adding a database for logging results, but due to the sheer size of the data, and my experience in those magnitudes, we will probably be looking for some SCADA partner for that.

Yes, and the team that could handle it.

I'm assuming (perhaps wrongly) that Embarcadero has similar reliance, and that both companies concluded that it would be a disaster if somebody else grabbed the technology.

Not that I know. Afaik they simple hoped to gain a market for their databasetools, while they bought an essentially profitable business, so it was not too big a financial drain.

So I think they bought more the channel (to customers) than the assets.

It's worth noting that a few other companies relied heavily on the TopSpeed tools, for example Psion (early PDAs) for whom (again IIRC) there were some special hacks in the linker.

I know there was some use in circles around Philips Medical too (startups). I was approached 2 or three times about helping out with installs and doing quick assesments of code (like how dos dependant it was).

Competition with Java is mainly at the educational level. As far as I can see, in companies, Java is mostly beholden beholden to the multinationals.

And lazarus for now doesn't really have much chance there, an occasional skunkworks excluded. You need to be really big profile to get through to those decision makers.

"Interesting startups" are always using stuff that is "hot". It's the nature of startups, specially ones that are perceived as creative. I usually ignore such messages, since such new trends are as likely to fail as to succeed.

> - the longtime neglect by Borland, due to financial mismanagement (siphoning > profits to the ALM stuff) and failed .NET adventures. ALM was doomed from the start,

I think the whole ALM vision was a desperate attempt of Borland to get some status as prefered vendor with a couple of multinationals, and then build on that.

I can't seriously imagine that they really thought this was one of those enterprise technologies that gets pushed downmarket to middle and small biz. Maybe some downspeced parts, but not the whole shebang, and not against enterprise prices.

but with .NET Borland was "damned if they did and damned if they didn't": if fewer compiler vendors had signed up for it then MS might have felt under more pressure to make it succeed.

Well yes. There is a touch of hindsight there. It probably was a conflict to the safe exploitation of a year-on-year decreasing business that might face a sharp edge down if .NET really took off, and the prospect of taking a risk and conquering a new market. But I think that is exactly what happened, the risky option was chosen by default (we have to do something, and what MS does is the next big thing), and not on merits for Borland's situation.

Added to that , maybe Borland saw the trend coming that a large portion of users would stop autoupgrading after D6/D7 coming.

Some observations (opinions which I mostly collected from discussions on Borland forums, and general IT press at the time):

First, there was the mistaken assumption that .NET would conquer all. All the press said this, the "visionaries" in the fora reiterated this time and time again and there was never much critique on this. Even though .NET was an instant hit, there was never much discussion if it would stop at 60%, 90% or 99% of the market.

Second, (and I then already warned for exactly what has happened) a win of .NET is not necessarily a win for everybody that jumps on the .NET bandwagon. MS, as platform host would come certainly come out on top.

Added to that was the fact that MS didn't really push the multiplatform agenda of .NET. There were no big partners to push this, no compatibility framework, just a not production ready side-project (Rotor).

If this was panned out more, the independant ISV's role there would have been much larger, since MS was unlikely to do more than token support for those, and the independant toolchain makers could build a case around multiplatform .NET support. However the big partner that would have laid down the missing infrastructure for this was missing. (Novell/Mono came only into view much later and never was MS partner in a way visible to developers, except in press releases)

Third, the big hit of .NET is mainly the serverside part (ASP.NET). Not surprising for a Java copy, since that is a Java stronghold too. Borland went for clientside .NET, that is way less succesful. Clientside .NET didn't fail altogether, but was definitely second rate, and fragmented (Delphi.NET, winforms 1/2, now WPF, which then got spun into a RIA thing etc)

Fourth, Delphi.NET was a compromise from the start, to not startle the users with compatibility breaks or multiple paralel toolchains (the Kylix/CLX fiasco was still fresh). However it was clear that Delphi.NET was a transition horse only. The native <-> .NET dual strategy couldn't be prolonged indefinitely. And when the hordes fail to migrate, this was not something Borland could keep up for long. Worse, they knew sooner or later they had to tell their Delphi.NET users to do a second painful migration from the transition horse to whatever .NET only product they had planned (never heard details about this btw, which scares me. No Prism doesn't count)

We do that, running as a bureau service and sending reports by mail 365x24. We monitor sites from Sweden to Turkey, apps in Delphi with PostgreSQL as backend.

Apart from things directly related to that, I'm considering using Lazarus+Postgres for a multi-level collaborative project. The point is that Postgres has a facility where it can notify each user that there's been a change, so it's comparatively easy to have a shared work queue. That's the sort of thing I mean by "database in an unexpected situation".

Always seemed to me that Clarion's star customer was Frito-Lay, or perhaps my informant was excessively into junk food :-)

And lazarus for now doesn't really have much chance there, an occasional skunkworks excluded. You need to be really big profile to get through to those decision makers.

Yes, I used to sell development tools to a lot of these people.

The specialist press is enormously influential in this industry, and frequently ill-informed.

Added to that was the fact that MS didn't really push the multiplatform agenda of .NET. There were no big partners to push this, no compatibility framework, just a not production ready side-project (Rotor).

I agree, but having Intel+AMD pushing x86 and then x64 and steamrollering the opposition didn't help.

Agreed. However there's obviously JavaScript which also existed in both client- and server-side implementations, and has survived in the former despite MS's hostility.

The really big thing though- and this affects .NET, Delphi, and a myriad of other tools and environments that have been and gone- is that it's impossible to play "catch up" with Microsoft. If you can carve a niche for yourself and keep them out then you have a chance, but once they get an inch ahead you're in the dust.