Software Development

Is the end nigh for Borland?


You know the whole rubbernecking thing, where people slow down near an accident so they can take a moment to look at the carnage? I kinda feel like I'm rubbernecking whenever I read about the slow decline of Borland, as in this little piece on Valleywag earlier today. Because everything that company has done lately seems to be a wreck, and like a spectacular car crash, I cannot turn away. I can only look and wonder how on earth the accident happened.

This will only expose how old I am, but I actually wrote a DOS Turbo Pascal application in my first job out of college. Borland tools were just head-and-shoulders better than anything Microsoft was churning out. At the time, Turbo Pascal was possibly the best development environment around. Compared to the command-line VAX compilers I'd used in college, this thing was a dream. It made me realize PCs were definitely going to rule the world, if they had development tools that much better than the stuff on mainframes and VAXes.

But then Borland stumbled trying to become a general-purpose software company churning out spreadsheets (Quatro Pro) and database software (Paradox). They spent a ton of time, money, and focus trying to beat Microsoft at Redmond's game instead of doubling-down on the one area -- software development tools, where they had a substantial lead.

The rest is history so well-known it's almost folklore. Microsoft crushed Borland (and later Corel) in the desktop applications business. And when Visual Studio came out, it marked the beginning of the end for Borland's development tools business. Those early releases weren't impressive, but in typical Microsoft fashion, each release got better and better until Visual Studio was just so much better than anything Borland was putting out. 

Yeah Microsoft may have cheated, seeing as how they were also building the OS that the dev environments were targeting. But I think Borland could have kept ahead of Redmond if they'd stayed focused. Look at Intuit -- Quicken is still the big player in personal financial software despite Microsoft's best efforts because, unlike Borland, Intuit stayed focused on their core competency.

Now even if Borland had stayed focused, they would still be faced today with the challenge of open-source software. I mean, who wants to pay money for a dev tool from Borland when Eclipse is free? But again, I think if they'd stayed focused, they might have found a way to navigate these treacherous waters. As I commented on the Valleywag piece, Red Hat has proven you can make money in open source -- you just have to find a niche that you can fill. Maybe Borland could have been a bigger player in the early Eclipse days and then turned around to sell value-added tools and support on top of Eclipse.

Instead Borland has stumbled badly. First, they turned down a cash offer for their development tools business, then they spun it out as CodeGear for a possible sale or IPO, but now it's getting sucked back in to Borland (notice that the logo on the CodeGear site says "CodeGear from Borland").

And while they've had some success with their other businesses, their messaging around this "application lifecycle management" track has often been lackluster and confusing. Ask 10 developers about Borland, and count how many know about the IDEs vs. how many know about the ALM products.

I reluctantly have to agree with the Valleywag posting at least when it comes to IDEs -- the latest moves regarding CodeGear make me think at some point they will have to jettison that business entirely. Ironically, perhaps they will open source it, and find a new life for the products that way.

What do you think?  Is it over for Borland?  What could they have done differently to still be relevant in the development tools space? 

23 comments
jslarochelle
jslarochelle

I actually wrote programs in Turbo Pascal 5.5 (or was it 5.1 ?). I think I still have the floppies. At the time the Borland tools were way much better than MS tools (Quick Pascal...). When Borland C++ 3.1 came out with the Turbo Vision application framework I was able to use this to write utilities for our in-house interpreter (a continuous online FTIR application scripting tool). I wrote a nice shell that was use to integrate the different operation that had to be performed to edit the configuration files and start different utilities. One interesting aspect of this application was that it would take almost no memory when spawning other application (remember DOS ~640K). The secret: a microcospic shell that I wrote using Tasm. Of course you could debug this using Turbo Debugger (ligth years ahead of the corresponding M$ tool). When the time came to write Windows application I did not hesitate and I used OWL (again much better than MFC). I wrote an application to control a laser using this and it was still working under Windows 2000 a couple of years ago. When Delphi came out I switch to that for utilities until C++ Builder came out. I still use C++ Builder on those rare occasion when I have to write little .EXE. The switch to C++ was because of the large amount of C and C++ code already written by other programmers in the company. It was just easier to integrate that with a C++ tool. Borland dying would really be sad for me. I don't think that having a monopol on commercial development tool will be good for Windows either. But I don't expect M$ to understand that. They just don't understand the importance of diversity and fair competition. JS

Tony Hopkinson
Tony Hopkinson

Went back to VAX and Fortran for a bit, couldn't get back to it. Took me two years to source a decent job and my employers nine months to source an acceptable candidate whose skills included Delphi. Tells me everything I need to know. Six Delphi Developers in our company, moving to C# and .net, neither us nor the company see sticking with it as a viable business or career option. Pity, but they screwed it up not me.

onyxps
onyxps

Either you don't know what you are writing about or "somebody" is paying you for deliberately spreading FUD. Shame on you !

Neon Samurai
Neon Samurai

I guess that's not really a good sign as the last I remember from them was the Borland C++ package though .NET seems to be the defacto Windows IDE now.

frank
frank

Yes, the end is nigh for Borland - in fact, the company is a rotting corpse, but because some aren't aware of it yet (?), they can be called a zombie, if you will. Borland should never have allowed Microsoft to hire away 35 of their key people, or at least find competent replacements. This has lead, 5, 6 years ago, for Delphi (Borland's flagship cash cow) to become a follower instead of a leader, but if you ask me, Delphi has been deliberately killed off in the 200 million USD cash deal with MSFT, where an agreement was made not to make Delphi cross-platform, not to provide native 64-bit functionality and not to provide Unicode-functionality out-of-the box. Basically, Borland sold their soul to Microsoft for two hundred measly million bucks, and the company has behaved like a headless chicken ever since. Delphi's current program manager, Nick Hodges, has always maintained that Unicode, 64-bit and cross-platform would be "bad business", so we expect no change. Moreover, Borland/CodeGear is heaping disaster upon disaster, lately by aquiring some Spanish dude's buggy PHP IDE and selling it off as "Delphi for PHP". Nothing better to spread the word that "Delphi is dead" than releasing a dud like that.

rharris
rharris

Remember Inprise? Borland has wanted to shed their development tools for years. Always thinking bigger, but never achieving big-ness seems to be their fate. Invariably, they will return to being the cool company that developers love, and will leapfrog the competition (open source included) in the development tools. And they will start making money again so that they can start thinking big...

kvo
kvo

Borland is guilty in current situation not lesser than MS! Instead of concentration on _multiplatform_, quality software development tools, they decided to become Windows-only company. One example: very nice 'TopSpeed' family of tools that they rejected to continue, I could remember many other similar decisions.

Mike Page
Mike Page

I've got 19 years of code written using Borland C/C++ compilers. This includes lots of non-ANSI VCL user interface code. It would be painful to switch to another development enviroment. The current Borland compiler (BDS 2006) works great. VCL makes doing the UI reasonably easy. So why switch? If Borland doesn't produce a 64 bit compiler in the next couple years then I'll have my answer.

mattohare
mattohare

I'd just learned Pascal on a VAX and went into the world to do some programming. I remember it like it was yesterday. I went down to the brand new Software Etc in the brand new Westlake Mall in Seattle, and got a copy of Turbo Pascal. (Then went back a day later to swap it for a box with the right floppy disks in it. I didn't have 3.5" yet.) Once I got it in and running, I took a crack at a simple program. Others loved TP, I knew I would too. Then came the 'differences.' I found TP to be so different from the ANSI Pascal I learned in college that I'd be as well to learn C. Then I went and got QuickC. Next chapter for me in the Borland story was with Paradox. R:Base was dying slowly. They had a better product than others on the market, but were too slow to get the new versions out. Again, I heard great things about Paradox. Spent a weekend converting all my databases. Then found that it would let me set up views that took edits but not save them back to the tables. I spent the next weekend converting to MS-Access 1.0. MS has a lot of issues, and Borland may out do MS sometimes. (Everyone knows it's not difficult to improve on MS a lot of the time. *chuckle*) It's like another Seattle thing, Starbucks. There are better coffee shops, but Starbucks is still a pretty high minimum standard.

jgnderitu
jgnderitu

In Africa we have a saying,"He who competes with an elephant in farting ends up bursting his ass.." But then again Borland could also regain focus and make headway where they have been flogged.

vv123
vv123

The Delphi manager, Nick Hodges is very incompetent. He speaks in the newsgroups(NG) and confuses and twists everything around to make himself look better. All of the huge threads in the NG are about how Nick has confused everyone and by saying one thing, but meaning another. That goes for SOX, upgrade path for Turbo Users, problems with software Activiation and the debacle with the PHP version. Since BDS2006 he keeps trying to split the same seed to grow more money out of it. He went from BDS2006 (which is a suite of 4 different languages including Delphi ), to a couple of Turbo versions of the same Delphi, which can't be upgraded to the full blow BDS2006, you most pay full price for that. He also created Delphi for PHP, which is nothing but some open-source PHP components surrounded by a bad IDE, again, no upgrade path for it and now Delphi 2007, which is NOT part of the BDS2006/2007 suite, it's NOT a Turbo product, but what it is, it's an entirely new product line for the 2007 Delphi!! Nick has seriously fragmented Delphi to the point no one knows which product to purchase for fear that it won't have an upgrade path and that the quality will be just as bad as all of the other multiple copycat clones of Delphi. If they can get rid of Nick and also hire a C++ Product Manager(they still don't have one), they may be able to pull into a close 2nd behind MS. But for me, they will never see another dime of my money. We are in that group the jumped ship. Half of our team (the C++ side) jumped ship after the terrible release of BDS2006 and the rest of the team jumped ship after we found out that Turbo Pro users don't have an upgrade path the BDS (even though Nick said we did). Buyer Beware!!

Tony Hopkinson
Tony Hopkinson

If they insist in competing directly with VS2005 and .NET. It's not Visual Basic, and it's not Visual C++, Anders and co saw to that. Stepping from Delphi to it has been little more than a syntactic hurdle, I dare say I had an easier time of it than those who went from VB6 to VB.net. I did n't not use VB because it was MS, I didn't use because it was crap compared to Delphi. VS2005 and C# and .NET compares very well. They still had chance on performance compared to any ,NET development, they threw it away. They need to get back to basics, if they don't want to end up dead. Codegear is a gimmick, and a cheap nasty one at that. Should have kept Kylix up to date as well.

RexWorld
RexWorld

I'll accept your criticism, but only if you spell out which of my statements constitute FUD. You can't just accuse somebody without providing an example or two. For example, if you could provide a recent example of a company switching from MS dev tools to Borland that would be a good counter-point. But everybody I talk to has exactly the opposite experience--anybody with a Borland-built app is actively porting, or considering a port, to the VS.NET IDE.

Tony Hopkinson
Tony Hopkinson

Paradox 2.0 for DOS (87). I used Turbo Pascal for utility exes. Started with Delphi 1.0 in 96. Still doing Delphi now and moving to C# .net in VS 2005. I'm afraid I've got to agree with everything he says. Delphi 7.0 was the best windows development environment ever. It's still as good technically as VS 2005 Then we got the shite that was 8, as for 2005 which I'm still on, it's the flakiest piece of rubbish, I've seen in 30 years of development. They went .Net, they still haven't, last I heard got, .NET 2.0. MS are now working on 3.5! What ever persuaded them to go down that route I will never know. If I was them I'd ditch .net, in fact ditch 32 bit, go for 64, start looking at 128, start looking a multi processor. They beat the crap out of of MS by being in front of them, now they are behind and can never catch up. I'd love to meet up with the half wits who rewrote the IDE in C++ .... The company is moving to .net, because competent experienced Delphi developers are hard to source. I'm keeping my hand in, it will be a useful niche skill, having .net as well makes me doubly valuable, as that the most likely move for companies in the same position. Get out while you can, short of a big change, market wise it's dead. :( They were never going to out produce, they were never going to out market, they dropped the quality of the IDE and the compiler which was their flagship in an attempt to do so. Stupid beyond belief. In Business the only reward for sitting on your laurels , is a bad smell and a green ring on your arse.

drrobert
drrobert

I too, have copies of Delphi 1..5, skipped 6 and just purchased 7. Agree that the current help documentation is not as good as older versions. As an aside, Mr/Ms Harris, please contact me at my private email drrobert@sturec.com -- Robert

degreed EE
degreed EE

As long as I can remember, the question, "Will Borland survive?" has been debated. Delphi version 1 was a death march project, whose success saved Borland. Yes, I was sad to see first Anders and later many more sign on with the Dark Side, but please pardon me if I don't get excited by the latest news and speculation. D7 and D2006 work very well, although the help is weaker than it used to be, but personally I don't need it that much anymore. The advice in their forums is still excellent, and we're way ahead of MS in that way.

davechampion
davechampion

Nick Hodges is a product manager that has a passion for the product he promotes. The Code Gear/Delphi community in general has great attachment and passion for the products under the RAD Studio umbrella. I see a coherent plan unfolding and better quality software released and more engagement with us the customers. Nobody is perfect. Blow off steam if you like but don't insult someone who is doing a good job.

jslarochelle
jslarochelle

... update Kylix. I think that having the capability to port native application with such a good tool was an asset for them (and us). JS

davechampion
davechampion

You might see the Borland ALM business merge with someone else. And CodeGear as a separate company for a while. I can't see the two businesses continue as they are, there is so much they disagree about. CodeGear seem to be doing sensible things. There is a coherence in their strategy. (i) Maximize revenue in the short term JBuilder 2007, Interbase, D2007 and D4PHP (ii) Release D2007 as a quality release with the help of trusted members of the community (iii) Release C++Builder 2007 soon (iv) Ramp upto a Studio product later on in the year that also embraces the later Dot.nets. Nobody is dead. The end of the world is not here yet.

davechampion
davechampion

I agree that a cross platform solution is an obvious way to make Delphi more attractive. I think it would strengthen Delphi2007/8.net. A tool to single source across Linux, Win32 and Dot.net - maybe even MacOS. If Kylix were revived it would be as a Delphi plugin to Eclipse since this would complement JBuilder (and Ruby when released). But there have been no signals from CodeGear in this direction. A more likely approach would be something similar to Simon Kissel's Cross Kylix. But this would need solid brass balls to action alongside the much needed Win64 and Unicode work.

Tony Hopkinson
Tony Hopkinson

Got two to three years to run on the existing codebase while we switch, may get D2007 just because of how bad 2005 is, but use will be kept to an absolute minimum. They've been working on .NET 2.0 for ages, they've had ages to do it. I'm looking at 3.0 and 3.5 for the rewrite. No way that will be done as I need it. I can't afford to wait for them, you can't and we can't. To beat MS they have to be better at something D2007 is not and never will be better at .net than VS.... That would require MS's cooperation, Borland need to do something else, such as 64 bit, parallel pascal etc.