What programming languages are you using to create your games?
Does anyone here use the Blitz products, or other game creation systems, to create shareware games?

I'm a game designer with a programming background (mostly in authoring systems, though) and I'm wondering what is the best alternative - invest an year or so in learning C++ or going with a simpler solution first and learn C++ only after getting some games on the market...

Thank you,

I started with Java and I'm going to stay with it too. But years ago when I was just a hobbyist I used BASIC, because it gets good results - fast.

Cas :)

Well, I took the long and hard road... started out with BASIC when I was very young, continued to PASCAL (anyone still remember? :D ) and on to C. But what really impressed me the most was the step towards an object-oriented language, first C++ then Java and later C#. (Others I know are usually not used for games or similar content) OO so changed my way of thinking, that it's hard to program something in C-only, like I have to do right now. But back to your thread: ;)

I don't think it takes a whole year to learn C++. Naturally, it all depends on your prerequisites, but if you have a programming background and are really determined to learn, then you should be able to write little games in a few months. The experience cannot be rushed however, and along with the more difficult concepts of a language will take time...

If I had to start programming games right now, I'd probably take Java. Maybe that's because I fell in love with the clean code produced with it (compared to my previous experiences with C++) but I believe it is easier to learn as well. Also, syntax is somewhat similar to C++, so you should be able to make the step. (You'll miss some things very much, though!) The only thing that might be difficult is that you have to program object-oriented. If you'd rather not do this all in one, then there is still C, the mother of all those languages... ;)

I took the long road too. Started with BASIC at 7 (fifteen now, heh). Played with VB until about 11, when I started trying to learn C++ and went running away with my tail between my legs (finally got a decent handle on it this year). I'm taking classes on Java next year (pity me).

If I was learning now, I'd start immediately with C++. If you're confused by it, bite the bullet and keep going. Damn the torpedoes and all that. :D

I'd jump right in with C++ and good game programming library, maybe Allegro - http://www.talula.demon.co.uk/allegro/ . On the tutorials page there's an article about setting it up for Dev-C++, a free IDE/compiler package. The demos should be a good jumping-off point for getting started with your own stuff and you can pick up the language as you go, using the demos and an online c++ reference (like http://www.cplusplus.com/ref/ ). not exactly a standard reference, but another good, free read on c++ is Thinking in C++ - http://www.mindview.net/Books/TICPP/ThinkingInCPP2e.html .

Im using 3D Gamestudio for the game im working on at the moment. Its language is similar to plain old C and it can be reasonably powerful. I decided to use a third party engine instead of C++ because of time constraints and i still have nightmares over compatibility problems :(

I dont think id use gamestudio for any other projects though since it has more than its fair share of problems and limitations. Im planning to write my next game in BlitzBasic since ive heard only good things about it.

My (shareware game-developing novice) advice:
Don't plan on learning a language or technology to make a game.
Plan on making a game and use the best technology you have at hand.

If you already know something, it's probably best to use that.

*** Rant Begins ***

My big project for the last week has been to choose a platform for my first game.

Like a lot of people here, I started at a young age with Basic. Pascal was the language I fell in love with, that taught me structured programming and the fun of null-pointer exceptions. But it wasn't meant to be.

I learned x86 assembly language, and I learned to use it inlie to optimise some routines.

I learned C and eventually C++.

I learned how to address the video card directly and write low-level line drawing routines. I learned how to make sound come out of my sound blaster before there were drivers or libraries I could use. Sound mixing, module playing. Then polygons, then with clipping and gauraud shading. Precomputed sine tables, matrix math. Mode X, fixed point, double buffer. Real mode, 32-bit addressing.

Win32. Oh, what an adjustment. DirectX 3 - 1000 lines to put a triangle on the screen. I liked my software renderer better.

But I had to get a job, so I learned VB, SQL, HTML, ASP.


In short, I've been programming for 15 years and I've seen a lot. As of yesterday, I was making my decision between using Macromedia Director and using C++ with SDL.



My plan is to make 3 games over the next 9 months. First, a space-shooter, then an adventure game and a puzzle game. These will be the best games I can do with 12 weeks per project.

In the next two years, I plan on implementing features in updated versions of these games that Director may not support.


The way I see it, the exact game I would make with 12 weeks each in Director will take me 6 months each in C++ & SDL.

In one day with SDL, I realized that SDL_Mixer was a completely seperate library to learn. So are SDL_Image and SDL_Net. Demos that work fine in GCC may be incompatible with VC6. I modified a demo so it would work compiled with VC6... mostly. It unexpectedly exits from time to time.

The Macromedia games will Just Work. Eventually, I plan to move to a C++ platform, but only when I can afford the support issues.


Regarding Allegro: It seemed kind of old-school to me. SDL appears to be a more modern alternative. Most importantly to me, SDL interfaces with OpenGL, for accelerated 2D and 3D.


My conclusion: I may prefer to program in C++, but it will be much cheaper and faster for me to build in Director a game of even higher quality.

In this case, quality means:
- Runs on all my potential customers machines without support problems
- Easy install
- Time spent on better graphics and gui rather than time spend deciding how to store and load graphics and develop or learn a gui library that works with SDL with no wysiwyg editor.


The extra $1000 US for Director is well worth the cost.



Other options I looked at:
- OpenGL - no, I want as large a possible audience for a simple 2d game, and I don't have time to deal with card/driver differences. (doesn't Just Work)
- DirectX - To big an install to bundle with a 2-10 meg downloadable game.
- Multimedia Fusion - not OO enough for me. (lingo is barely OO enough for me)
- Jamagic - too immature: I won't be an early adopter on my first game
_ GLUT / GLFW - I want to spend more time programming game code, not engine code at this point.


*** Sidebar ***

On a sidebar, my daughter is now 7 and that is how old I was when I was first exposed to Basic on my sister's TI 99/4A. I just gave her her first computer for her birthday, and I'm planning on introducing her to the idea of making computers do things of your own imagination.

I think flash and director are too complicated. I'm looking for something simple like logo or TI-basic. Maybe even hypercard, as much as I despised it. I'm considering trying Multimedia Fusion.

Originally posted by mg_mchenry
- DirectX - To big an install to bundle with a 2-10 meg downloadable game.
- Multimedia Fusion - not OO enough for me. (lingo is barely OO enough for me)


DirectX: Almost everyone already has the DirectX files. If you're worried, use an older DirectX like DX6 - IIRC, Win98 came with DX6 or close to that preloaded.

As for OO, why does EVERYONE say "not OO enough for me"? Do you find programming without the almighty class too difficult? Perhaps it's just me (I learned C first), but structured-type programming works quite well.

Thank you all for the info, seems like I need to clarify a bit my position :)

I want to create a set of 2d games for low-end Wintel machines.
I need mouse control, OnMouseOver-type events, file formatting, minimal graphical effects (2D animations)...

I already used the Macromedia products on a large array of products (mostly multimedia presentations/encyclopedias, no games) and I always found them a tad slow and having some compatibility problems, but that was before 1999/2000 so I'm not up-to-date with the latest versions.

My other alternatives are:
- the Blitz products - namely, BlitzPlus and (for a different project) Blitz3D. Advantages: competitive time-to-market ratio. Disadvantages: I'm not sure if I can do everything I need...
- DJGPP/Allegro. Advantages: flexibility. Disadvantages: learning curve :)

If you know of other game creation systems, if you want to share your experience in developing games with such gcs, please make this thread grow.

FYI, you don't need to use DJGPP with Allegro, as it has supported Win32 compilers for a while now. Also, if you want OpenGL support, there is the AllegroGL add-on library.

For me, it was ZX Spectrum BASIC from age 5 to 11, gwbasic and QuickBasic to age 13, and C++ since. I'm 22 now and still learning.

C++ may have a hard learning curve, but hey, I learned it at age 13 on my own, with a (bad) book. Once you get used to it and become a "power user", you can do anything. Its raw power is awesome.

Other languages promise ease, portability, and the cure for cancer. Let me quote a conversation between a couple of friends...

A Programmer's strength flows from the Source. But beware of the dark side. [...] If once you start down the dark path, forever will it dominate your destiny, consume you it will.

Is the dark side stronger?

No...no...no. Quicker, easier, more seductive.

So, to all Java and C# developers out there : don't be too proud of this technological terror you've constructed. The ability to code a dialog in 15 seconds is insignificant next to the power of the Source...

As I said, I learned Basic and Pascal before I learned OO programming. I did structured programming for 10 years before I went to OO.

It took a long time for me to see the light, but I will not go back.

I think game programming is about the best case for OO programming I've come across. I've had Business Objects cause as many problems as they solve.


But a game entity - man is that an object. It's AI is another separate class. Changing the way that thing behaves is now as easy as it gets.

OO gets rid of huge blocks of if statements and switches.


yeah, you can make any software imaginable without oo.

I just prefer

Loop through game entities
entitity.move()
End Loop

to

Loop through game entities
If entity = packman then
packman movement ...
packman movement ...
packman movement ...
Else if entity = ghost then
think about ghost stuff ...
ghost movement ...
End if
End Loop


I'm not saying this is necessarily the way I would program this, I'm just trying to make a simple example.

Are you saying that if the death star wasn't coded with some RAD tool it wouldn't have been so easily blown up?

Because I always saw it as a C structure fallen victim to an unforseen buffer overrun error.

mg_mchenry : I find your lack of faith disturbing...

i took the short way and learned c++ at 12 1/2. i still use it a year later :) i recommend the book that diragor posted a link to (thinking in c++) if you are going to learn c++. i agree with mg_mchenry, you have to use too many if and thens in procedural programming (IMO). plus your update function gets huge (you don't have object->update's). but then maybe thats cuz i never really programmed in a full procedural language.

That's one of the beauties of C++ : it doesn't force you to do anything. Use as many ifs and gotos as you find necessary. Be as OO or as procedural as you like. Mix paradigms any way you like. Throw in some generic programming to the mix.

I prefer a practical language over a "pure" language designed to show a concept.

Please, no offense intended to the pre-teen and teenage programmers.

I'm wondering how early someone learns to program and when certain concepts can be introduced.

And by someone, I mean the type of someone who gets 99th percentile on standardized tests and might be the type of person that someday programs professionally. An above-average person with engineering smarts. The type of people I had a hard time finding growing up, but seem to be here in abundance.

When I was 5-7, I could do little more than copy programs out of books, and maybe change some of the numbers that had to do with how long or how often something happened. But no one else in my family could do that either.

I didn't have a computer again till I was 10/11. Now I understood program flow, subroutines, and basic logic and could make simple games.

I didn't do structured procedural programming until I was 13. At 15, being proficient at Pascal, I didn't understand the need for OOP fully until I was about 20.



How do you manage to learn C++ and OOD at 12 years old? I guess it's easier with the internet and all of the modern (and free) tools at your disposal. But can you really learn OO programming before you learn good procedural programming? And can you really learn procedural programming before you write programs with no procedure and realize how bad an idea that is?

I'm struggling trying to figure out what to show my 7-year-old daughter this summer.


Hey, I just thought of an idea for a game... Solve puzzles with logo-like turtle tools...

Originally posted by Eddie
Thank you all for the info, seems like I need to clarify a bit my position :)

I want to create a set of 2d games for low-end Wintel machines.
I need mouse control, OnMouseOver-type events, file formatting, minimal graphical effects (2D animations)...

I already used the Macromedia products on a large array of products (mostly multimedia presentations/encyclopedias, no games) and I always found them a tad slow and having some compatibility problems, but that was before 1999/2000 so I'm not up-to-date with the latest versions.

My other alternatives are:
- the Blitz products - namely, BlitzPlus and (for a different project) Blitz3D. Advantages: competitive time-to-market ratio. Disadvantages: I'm not sure if I can do everything I need...
- DJGPP/Allegro. Advantages: flexibility. Disadvantages: learning curve :)

If you know of other game creation systems, if you want to share your experience in developing games with such gcs, please make this thread grow.

I would learn something along the lines of VB if you want to use a RAD language...at least it has buisness applications. =) For 2D animations and the mouse stuff it should be fine but the file formatting its an inferior choice to C++.

Anyway, you people are having way tooo much fun with Star Wars =)

@mg_mchenry

I learned C++ at 13 and managed to get a 5 on the APT after not touching it for 3 years. And I did have some semblance of a procedure but not a good 'professional' approach to it. However, I doubt the APT are a good indicator of anything but its the best example I can give you at the moment..

I say teach her a scripting language or a BASIC one =) Its probably the easiest and you can do more stuff with less experience.

Originally posted by Eddie
I already used the Macromedia products on a large array of products (mostly multimedia presentations/encyclopedias, no games) and I always found them a tad slow and having some compatibility problems, but that was before 1999/2000 so I'm not up-to-date with the latest versions.


If you already know the Macromedia products, I highly recommend looking into using the newer version. It's a professional-grade product.

For various reasons, it may not be right for you. At least two of the games I'm making now in Director, I plan on porting to C++ when I can afford it.

Other projects, I may always do in Director.

Originally posted by grbrg

I don't think it takes a whole year to learn C++.

Yep, it takes many many years. :) I started programming in C++ about 4 years ago when I started college (graduating in August wohooo!). The more of the language I learn, the more I see how much I don't know yet. I started programming when I was about 13 if I remember correctly. The first language I got into was QBASIC. Ahh, those were the days. :) Naturally I soon realized the limitations of that language, so I switched to Visual Basic not long thereafter.

In High School they taught us Pascal for one year, then they switched to C++, but I didn't start using it until after I left High School; I thought it was too complicated at the time. Anyway, after I started programming with C++ I never looked back. :) Though when I get into 3D games I'm leaning strongly towards Blitz3D because it will speed up development time tremendously. The only problem with BlitzBasic is that it's too simple (not strongly typed etc) and not object oriented. Lack of OOP makes writing big programs a lot harder. If you're starting out and want to make games though you should take a look at BlitzBasic (http://blitzbasic.com) as it'll be a lot easier to start with than C++ believe me. :)

Originally posted by mg_mchenry
How do you manage to learn C++ and OOD at 12 years old?
In my case, I said C++, not OOP/OOD. There are lots of concepts that you need to let mature inside yourself, until one day enlightenment reaches you and you "get" it. At 13 I didn't OO programming at all (but I did use pointers, which for some strange reason seem to scare people).

I experienced this enlightenment experience later, when I learnt Prolog at my university. It's a new paradigm, so you try to program using the paradigm you're used to, but with Prolog syntax. Most of my classmates didn't get past that stage. But it's wonderful when you finally free your mind and start thinking in the new paradigm - the code starts to flow.

Same with OO. I started doing structured programing with a C++ compiler. It was later that I started using C++ as C++. But when I see stuff like the Boost library I still feel like a newbie...

[b]I don't think it takes a whole year to learn C++.[b]

Learning? No, you can write C++ which successfully compiles in about a week.

Grokking (http://www.ibiblio.org/lou/old/ball/chronicle/jargon.grok.html) C++? That takes years and years. I'm in the 9th year of the journey...

This is one of the things that bugs me about the C++ situation. I don't think C++ takes so long to learn.

Imagine someone who has programmed in C and Java for some time, how long does it take that person to learn C++? Not long.

But:

How long does it take to learn your particular compiler?
What about the pre-processor?
Platform API (e.g. Win32)
Processor architecture
API's like GDI, OpenGL, D3D, etc.
Why doesn't it work when I do x?
Why doesn't it work on x's computer?
Installer. IDE, Debugger, threading.


You have to know so many things when you use C++ that don't have anything to do with C++ itself.

This, of course, is why we have Java. I want Java without the virtual machine, and it seems this is now possible.

Originally posted by ggambett
Learning? No, you can write C++ which successfully compiles in about a week.

Grokking (http://www.ibiblio.org/lou/old/ball/chronicle/jargon.grok.html) C++? That takes years and years. I'm in the 9th year of the journey...

I'd go as far as to say that the learning never ends... The question as I understood it was when a programmer reaches a level when he or she can make a little game.

OOP or not to? Well, that's a discussion that will probably never be concluded. OO will be one of the major concepts and there is no denying that. On the other hand:

C-Programmer: "If you kill me I will become more powerful than you can imagine." ;)

Originally posted by grbrg
C-Programmer: "If you kill me I will become more powerful than you can imagine." ;)

Yes, through a cleverly placed memory leak, your code will retain control of a part of your memory until the system is restarted. Better to let you de-initialize on your own.

In "the mythical man month", Brooks talks about an order of magnitude difference between junior developers and skilled developers. Someone may learn C++ in a year, may get their feet wet with platform APIs and various other things, but to become a skilled developer takes a lot of time.

Learning multiple languages also helps, I think. I had been using C++ for years, but never really grokked OOP until I used Objective-C. Then I used Java and realized that it's possible to be take OOP too far.



That's one of the beauties of C++ : it doesn't force you to do anything. Use as many ifs and gotos as you find necessary. Be as OO or as procedural as you like. Mix paradigms any way you like. Throw in some generic programming to the mix.


I agree 100%. Python is another language that I felt gave the same kind flexibility.

For me, Python was the eye-opener. As I said, my ideal language would be a mixture of C++, Python and Prolog. I miss Python features like built-in lists and hashes (Java fans : built in the syntax, not in class libraries), functions as first-class objects, map, filter and reduce, and many others...

Delphi is a great choice for several reasons..

1. when your done you can port to Linux.

2. its makes lighting fast Bin code.

3. its easy to integrate other languages into it if you feel more comfortable with them

4. its very object oriented , but if you wans to do procedural programming you can.

5. see reason 2.

Originally posted by Eddie
What programming languages are you using to create your games?
Does anyone here use the Blitz products, or other game creation systems, to create shareware games?

I'm a game designer with a programming background (mostly in authoring systems, though) and I'm wondering what is the best alternative - invest an year or so in learning C++ or going with a simpler solution first and learn C++ only after getting some games on the market...


Some personal thoughts because I recently dealt with this. If your game is small and doesn't need the performance boost I don't see the benefit of not going with something more RAD oriented like Blitz or Director. Both are great products for what they can do. Personally, I don't like coding in Director with lingo (the language). Maybe I'm strange, it just feels wrong. :) On the other hand Blitz is a Basic variant and it's familiar. I still prefer C/C++ (I use OOP when I think it's appropriate and procedural otherwise) but the longer development time isn't worth it for smaller games. Once I move up to my "Big Bad RPG" (which will be the best in the world of course) I'm sure I will move up from Blitz3D to C++ and probably a graphics library like Fastgraph or Power Render. But it takes time to work your way up there. I read in "Game Design Secrets of the Sages" that you should rate the difficuly to develop your game on a scale of 1 to 10. Start at 1 or 2 for your first game. For every game after that, don't make a game more than 2 points harder than your last one. After all my failed massive game projects, this sounds like pretty good advice. :) Also my first "real" game is around a 2 and it's 90% complete!

Josh

I went backwards.

I have been programming C++ for a long time, and have made many commercial CD-ROM and downloadable games with it. I switched over to Blitz3D because I firmly believe that I could do the same exact game in either, so I might as well get it done faster!

I sure miss the Visual C++ debugger though! (classes too :D)

C++ developer here. Coming up with clever OOP designs has replaced the fun and challenge that used to come from coming up with clever tricks to get your DOS game to draw quickly (especially when SVGA came around, 1024x768x16 on a 486 running smooth :D). C++ is rarely, if ever, used to its full potential. Things like the STL provide a wealth of functionality, as well as a lesson in what can be done if C++'s features are actually utilized. Now if MS would care to provide a standards compliant compiler (disclaimer: haven't tried .NET 2K3 yet), C++ would be an even happier world.

Originally posted by Mike Boeh
I sure miss the Visual C++ debugger though! (classes too :D)

You'll come crawling back eventually. They always do. :p

Michael Sikora
Guardian Light Studios (http://www.guardianlightstudios.com)

If it the blitz debugger were any good, I'd try blitz, too, even though I really don't like the BASIC syntax. Until the blitz people fix their debugger, I'll stick with C++.

RedNeck
Delphi is a great choice for several reasons..

1. when your done you can port to Linux.

2. its makes lighting fast Bin code.

3. its easy to integrate other languages into it if you feel more comfortable with them

4. its very object oriented , but if you wans to do procedural programming you can.

5. see reason 2.

One thing that really annoyed me while working with Delphi (5 years as system developer) is that I had to write := instead of = and ' instead of " and please don´t mention begin/end
Then you don´t have the keyword static (I like that one).
I could go on and on with things I don´t like about Delphi, but I won´t... :)

Delphi is very good for database applications, but I don´t think that I would ever use it for anything else.

The editor in Delphi is the best I have ever used, everytime you press a key the editor seems to know what you´re thinking, I can´t say the same about Visual C++.

Oh yeah, one more thing that I think is quite stupid with Delphi, is that you can create an instance of an abstract class, pretty dangerous IMO... (hopefully they have changed that in Delphi 7)

About your number 3, if someone feel more comfortable with another language, why use Delphi at all? ;)

I might have to jump ship, as much as I enjoy the environment.

I started doing perfomance testing yesterday.

You see, I'm using Director in a very unusual way: I create and destroy my sprites every frame because you can't set the z-sort value of a sprite - it's dependent on it's sprite number, so I re-sort the sprites every frame.

It seems all this sprite overhead is unneeded when I can just blt the images to the stage instead. I decided not to use sprites at all, and just copy the images.

Then came the testing. With an early version of my game, only a few AI, and maybe 25% of the script processing that will need to be in the full game and maybe 33% of the sprites on screen and I was getting 110 FPS.

Good, right? Not really. My target rate is only 30FPS on an unaccelerated 200Mhz? I only got 110 on my 2Ghz dev machine. I tried the game on a 350Mhz machine and only got 25fps. This won't be nearly fast enough for the completed game.

I tweaked and tweaked getting a few fps here or there, even taking out things that the game needed. It never got too fast.



Then I went back and tested how my original sprite-based method was performing. 188 frames per second. HA! Even just clearing the screen every frame, I only get 173FPS. Or not clearing the screen and just blting the images, I only get 142fps.


So, I was completely wrong! Director sprites are faster, even if you create them every frame.


But my problem is that I need to reach around 500fps on my dev machine for the game performance to be acceptable on the p200 for the finished game.


I'm testing some code that will allow me to not re-create the sprites every frame, only when needed. I will be suprised if this gets the frame rate high enough.

I'm also testing displaying the 2D images through Director's 3D engine to see if it's any faster.


But it looks like my idea of a simple game is still to resource-intensive for Director to handle. I may be forced to bite the bullet and just do it in C++.


One interesting thing I found in my testing is that Director parses and interprets comments in the script every time I go over it. If I used a big if statement to ignore the sprite display code that gets called 100+ times per frame, I got 500fps. But if I commented out the whole block, I got 450fps.

The comments take up processing time if they are inside your function. I never would have expected this. I won't be putting comments inside major loops any more with Lingo, or possibly other interpretted languages.

You see, I'm using Director in a very unusual way: I create and destroy my sprites every frame because you can't set the z-sort value of a sprite - it's dependent on it's sprite number, so I re-sort the sprites every frame.
What version of Director? We have recently finished a soccer game set to go online soon in which we change the z ordering of the sprites every frame. So it is possible, at least with 8.5, and at least with flash sprites.

I'm using 8.5. I've been using Lingo now for 5 years... maybe I've missed something along the way.

But if I have a sprite in a channel, the channel number determines the z-sort. Right?

So I rebuild the sprites every frame. How can you change the z-value?

It's far more possible that I am missing something. I'm a C++ guy and I just hacked two projects using Director.

That said... what we did was changing the locZ property of the sprites. Again, this might be useless, I'm no expert with Director.

To add to RedNeck's reasons Delphi is great:

6. Write as much code in as many units as you want, compilation will still be almost instantaneous.

Wow, there is a LocZ property. This is some bad documentation.

You can't look under "sprite object model" and find all of the sprite methods and properties.

The closest thing they have is lingo-by-feature. If you look under sprites, there is no mention of locZ. The only way you find it is if you read through every property alphabetically and stumble across it.


Grrr. Hey, thanks for the tip. There's always something to learn, I guess.

I agree. Director's documentation is pathetic. Worse than Lingo. Worse than that ActionScript interpreter which doesn't show any warnings if you call a non-existant method of a non-existant class :(

Originally posted by Diragor
To add to RedNeck's reasons Delphi is great:

6. Write as much code in as many units as you want, compilation will still be almost instantaneous.

Yah, Delphi is great when it comes time to build your project.

Both Artifact (client-side) (http://www.samugames.com/artifact) and The Journal (http://www.davidrm.com/thejournal/) were designed and built in Delphi.

Both projects are pretty big. Artifact's build compiles about 160K lines of source, The Journal's about 220K lines. Takes about the same amount of time either way (less than 1 minute).

Delphi is adequate to any programming task you want to put it, including games.

I used to routinely switch between C/C++, Pascal, a proprietary form of Basic, and even Unix shell script, all within a given day. Syntax differences are hardly a reason to like or dislike anything.

-David

DavidRM
I used to routinely switch between C/C++, Pascal, a proprietary form of Basic, and even Unix shell script, all within a given day. Syntax differences are hardly a reason to like or dislike anything.

Switching between different languages or learning a new one when you already know how to program is no problem, but I don´t agree with you on that last part.
Syntax differences is part of what makes a language and even if you get used to it, you don´t have to like it.
It´s all a matter of taste, so we better leave it there.

One thing that was great with Delphi was the fact that you had all the sourcecode for the VCL library (Enterprise version only I think), when there was some bugs, you could fix them instead of waiting for Borland/Inprise to release a SP.

If it the blitz debugger were any good, I'd try blitz, too, even though I really don't like the BASIC syntax. Until the blitz people fix their debugger, I'll stick with C++.

There are some 3rd part IDE for it, one is blitzview and the other Protean IDE (for this one you have to pay).

Surely having to buy a separate IDE is a bit disappointing...

About the question speed/reliability: the speed is nice, I'd say only 30% slower than a C program (it depends on many things of course).
Reliability: if I manage to finish my soccer management game, you can be sure you can program BIG project in it. I must say it is not the perfect language for big projects, because it lack a bit of OOP programming. You CAN do those, but it is better to do smaller games with blitz... unless you want a good headache :p