TUI/CUI
Ah...the good old days of TurboVision!
Embarcadero has released RAD Studio 10, including Delphi 10 and C++ Builder 10, a suite of development tools for Windows, Mac and mobile platforms. A new RAD Studio is a rather frequent occurrence. RAD Studio XE8 was released in April this year. The price is a hefty £3,612.60, plus VAT, for new users or £2,408.40 to upgrade. …
"Our latest non-legacy code won’t compile with the latest C++ CLANG compiler,” said DevExpress CTO Julian Bucknall"
Well Mr Bucknall, either the Clang compiler is broken (which is pretty damn unlikely given its usage), or, your devs need to start learning to code in standard C++ and not some ancient Borland dialect.
Hmm, wonder which it could be...
I actually used the CP/M version for a while on an Amstrad CPC.
Apparently it's still going strong (or was a couple of years ago).
And now I'm using VS 2015 and C#. Lawdy - d'ya think I'm getting old?
"Using left shift for multiplication is fine so long as you know whether your operand is signed or unsigned beforehand".
Very true. And can you guess what this was...? :-)
Other compilers at the time (I think I ended up switching to some GCC derivative, but I may be mistaken) produced the appropriate code.
The problem is not really the price, is what you buy at that price - it's not really worth it, the price/features and price/quality ratio is really too high.
For example, the GUI still relies on skins and "imitations" of the real GUIs (but the standard Windows one), and under the hood there's very little or no security in the communication/remoting frameworks. Often, to develop really good applications, you still need 3rd party libraries, which add to the price.
Some kind of developments, i.e. Windows drivers, are not supported, you need Visual Studio if you need them.
Moreover now bug fixes are available only if you buy the subscription, and old versions are barely mantained, or not mantained at all. With a new release every five/six months, working on complex applications whose development may take several months becomes a nightmare.
If you want to sell something at $3,000, people need to perceive they really get something in return...
Ironic, innit? IIRC Turbo Pacal was one of the first cheap IDEs on DOS, at a time when the MS counterpart (its C compiler?) was going for $500, a fair bit of dough in 1990 money.
$3000+ freezes out any beginner/hobbyist use. It might be worth it for a pro shop, if it made your guys massively productive - which you say ain't the case. But it will also severely restrict your hiring pool of young programmers if the language/framework is exotically unusual.
Ain't worth the hassle.
I've used most of the Turbo- products and am still converting databases from. My initial stab at Philippe Kahn's technologies was probably Turbo-Pascal but I enthusiastically embraced most of the others. I never understood what killed that endeavor other than a Microsoft knee-capping.
Most companies now understand that to get market share they need to get developers and trial users on board. Small opportunity windows like 15 days don't do any good. Limit the ability to build huge projects; limit the number of simultaneous developers (2?); limit distributing production products (1?). But don't cut off your afficiandos - the ones that will test your software, the ones that will blog and proselytize.
"I never understood what killed that endeavor other than a Microsoft knee-capping." Part of the problem was that Phillip Khan thought it far more interesting to play jazz than to run Borland. Yes there was intense competition from Microsoft but Microsoft didn't tell Kahn to lose interest in his company
I guess in a way we're still hearing the muffled screams of their demise. Do people still use Delphi? To be fair I liked Borland's little SQL server back in the day. Well as much as one can like a product that is stuffed down their throat. Still, it was a viable alternative to MS, etc for a small shop. Iirc MS SQL couldn't even do record level locking at the time without padding out your table schema to whatever pagesize. Take away the datatypes and you'd think MS wrote the first NoSQL database decades ago.
Borland C++ goes way back..in 1992 also targeted Windows, using the Object Windows Library (OWL) framework which many developers regarded as superior to the Microsoft Foundation Classes (MFC).
Borland was also offering a RAD C++ development environment for a long time. Borland Builder 1.0 was released in 1997. They did it in a very clever way (probably thanks to Anders Hejlsberg who is a personal hero of mine). They just wrapped the Delphi VCL and made it linkable. It meant we could use nearly all the existing Delphi components and even write components for Delphi developers.
I've often thought it was a bit sad that MFC remained the dominant C++ framework for so long. Well..I also chuckle a bit when I think of all those C++ developers fighting MFC when I was enjoying the benefits of the VCL.
The only criticism you could hold against C++ VCL was that it required a couple of none standard extensions to support the event model and properties.
I enjoyed using Borland's tools. I was even a member of their TeamB for a couple of years until a takeover at work dragged me over to Visual Studio. But hey - at least I still get to play with Anders' creations :)
I earned my first PC revenue using a Turbo Pascal application. I must have bought (and still have) copies of most Borland products over the years. Prologue was a bit of a miss but Paradox worked well for the price. I also used Delphi right up until CodeGear took over and still use it on legacy projects.
However, since Embarcadero took over, the price has risen in leaps and bounds to the point where it is not viable for small companies like mine. When a PC would cost you £1000, Delphi was £250. A PC today costs £500 and Studio is now £3,600.
Last year I switched to FreePascal and Lazarus. I moved my Delphi projects across to Lazarus but it was worth the effort as there is the bonus of Lazarus being cross platform; something Borland never really perfected with Kylix.
So, bye-bye Borland and thanks too to Philippe Kahn for his vision.
Well said Sir!
The company are doing an awful lot of price gouging these days. I guess less and less people are using their products. sort of ever decreasing circle really.
I stopped using their stuff when they moved to a License server. After 3 weeks, it just stopped worknig and they would'nt let me setup a new one.
moved to Lazarus and got on with development rather than spend time forever upgrading after shelling out lots and lots of dosh for the pain in doing so.
I remember buying Borland C++ for OS/2 and running into horribly frustrating runtime bugs with their port. They released 2 or 3 major patches and I swear that the situation became even worse. Things like threading APIs went from glitchy to crash-your-machine-non-functional. Then they discontinued support altogether.
I never trusted Borland after that and it was very frustrating considering I scrimped and saved from my meagre post-University bank account to buy their software. Fortunately I switched to IBM CSet++ and despite being command-line it worked as advertised.
I used JBuilder for a bit after that but was happy when Eclipse appeared and I didn't have to any more.
What was so amazing about the original Turbo Pascal was how accessible it was for such a powerful tool. I think it was £45 plus a bit more for the libraries. I had become restricted by MS Basic on a DOS machine but as soon as I switched to Turbo Pascal I was instantly productive.
I feel that Borland lost it's appeal for me with Delpi. Yes it was a good tool but for the type of business applications I was producing MS Access was far faster and had a better database back end.
Now the tool is super expensive unless you are creating a major program with a very big budget.
Agreed. I was using Delphi solidly from V1 in 95 up till version 7. In fact I was still using V7 until recently as the IDEs after that were too bloated and buggy for serious use IMHO.
Yes completely crazy how Embercadero have lost touch with reality. I remember you could buy Delphi 'pro' for around £300 which gave you all you need. If you needed enterprise DB access then it was cheaper to buy a component set from a 3rd party that would talk to your DB rather than pay for an enterprise version of Delphi. Then ADO came along and Delphi pro was good for anything.
Not only has their market share declined the price is going up whilst M$ are giving more away for free and there are plenty of free tools available for dev work, especially for mobile and web.
Crackers!
It started back in Delphi 4 days. In fact, Kylix was Borland's answer to the budding project, in an attempt to preempt it. However, when the developers saw Kylix's going nowhere, Lazarus was revived. A very useful little gem, if you ask me, even though the IDE could use more polish.
Wow, is that what an IDE looks like these days? I remember buying Turbo Pascal 5 in 1988 (came on 5.25 and 3.5 inch floppies in the box), then Turbo C shortly afterwards. I was a hobbyist programmer, so it was all trial and error (mainly error), fun stuff. Gave up when the Windows IDEs appeared, and saw how damn complicated (to my old, text-mode eyes) it was just to open a window! I was still getting junk mail from Borland in about 2001! Just for the pros now, I suppose. <sigh>
£3,612.60, plus VAT - I rather suspect that they're not getting many new users these days. They're probably mainly milking the users with an existing code base which they can't justify porting to use new libraries. It's an understandable business decision from both perspectives, but it leads to ever-increasing prices to squeeze out that last bit of cash from each remaining customer.
I used the original TurboPascal 3 on MSDOS, Turbopascal 4, and Turbo C back when these were pretty good products at what at the time were very low prices. They were seen as revolutionary in their day, as they brought quite good compilers (and later IDEs) at very reasonable prices to the masses. When Turbo Pascal originally came out, it was mainly competing against interpreted Basic (with GOTOs and line numbers). Turbo Pascal provided a very cheap product with compile times that were so fast as to give a compile/test cycle that was as good as from an interpreter.
Turbo Pascal started off as Compass Pascal from Denmark. Borland bought it and gave it the sort of mass market access which was difficult to get in those pre-Internet days. It came with a very good book, perhaps one of the best examples of documentation to be found in those days. There was no copy protection, but the book which came with the compiler was so good that it was worth buying (about $50 I think) just for that, making piracy not really worth the effort.
I used Turbo Pascal and Turbo C professionally and was quite satisfied with them. If you worked in the test and instrument field they were the most widely supported options for software development, with Microsoft's compiler trailing behind in support by vendors and pretty much everyone else being ignored by the hardware vendors (NI, KM, HP, etc.).
These days it's pretty difficult to make a living from selling mass market development tools when there is GCC and lately LLVM as compilers, and a million and one editors/IDEs available for free (I'm using Geany these days, which I'm quite happy with). A very big chunk of the market which Turbo Pascal originally addressed is also filled these days by Python, which comes "batteries included" and has a huge selection of free third party libraries available which do just about anything you can imagine.
Turbo Pascal was considered to be revolutionary it its day, but it's a very different world now.
You cannot compare those options. The RAD Studio, Delphi or C++ Builder is an offering that serves you well as a whole or not. As mentioned by others in this stream alternatives do exist for years now and those are well known. So an alternative offering must provide something different specialized but complete. That somehow seals the faith of all those who do not reside in the customer group to be addressed at the moment. Price is not an objective valuation or measurement. It's simply a compromise that can be found. The price of every good goes up in the end. Price can also express scarcity. I tend to think that EMB is thinking in terms of software as an asset.
Similar to any other tool RAD Studio does not fit into the early stages of utilizing a technology since especially in IT development does happen in a more 'communist' fashion in a traditional sense. What people call third-party in this what others call division of labor which is not cheaper but more flexible. I don't know in how far it is possible to industrialize IT especially software development from a more traditional view in general. I doubt that. What never did work and still does not work is 'communism' in the IT department and division of labor in the production process when a company is not aiming at IT but real world goods. Communism in a sense of getting the most out of ones own limited resource pool. Of course there is a will and a wish to extend the resource pool called knowledge - it's obvious why. As long as people are in the position to charge the customer for gaining experience everything is fine since the turnover is almost the profit. In the world of division of labor that's called expensive.
Today this resource pool can be extended by picking up source code from the neighbors garden called open source repository.
We all together are heavily working on providing both the commercial alternatives as well as the open-source route.
Embarcadero are not the bad guys that have taken away the cheap Pascal or C++ IDEs they simply made the experience at the time of Delphi or C++ Builder 2006 that offering a cheap or free alternative simply does not offer any advantage to them and also the number of users did not increase according to them or what was called Codegear. Not talking about why ...
Doing everything on one's own as an ideal. In the past when IDEs came into the focus (kind of Powerbuilder for example) those environments were signaling the beginning of the end of the underlying technology. That will not happen so soon for all the environments addressed, which puts a different light on the evolution and priorities that have to be fulfilled beside the new things offered.
The price is little high but in general you get discounts. From the pricing perspective I personally assume the vendor does not address the hobbyist at all (except from the starter). In USD but discounted even the initial offering does come at an acceptable price but for U.K. the price is heavy. The service agreement is pretty cheap. I owned E/R Studio for a long time and the service contract was cheap. Just the initial price is high. If you are a purchase manager in a company and get awarded for discounts :). Many companies are still bound to products that do offer service contracts and support mainly due to certification. Still the same story...
I personally tend to agree with LDS that when it comes to security the communist approach combined with materials picked in the neighbors garden seems to be the better alternative at the moment.
In order to address Hobbyists you have access to Appwave which is free on Android.
I personally like the RAD Studio since it serves my descent purposes well but RAD Studio is not an IDE that is focusing on communism as the major development strategy.
Some of the comments here are a bit misleading, mostly on price. That's for the top-of-the-line "Architect" edition. Normal Pro versions are, for a brand new license with all languages and platforms, less than half that. And if you only want one language - C++ Builder, for example - it's a third. And then if you are a hobbyist programmer or a small dev team, there's AppMethod, which is the cheaper edition, priced very nicely, as a per-month subscription of $25.
The Architect edition is comparable to Visual Studio, which is 5500 pounds for their top edition! So, especially given you can use it for $25 / month, I'm not sure complaints about price are really justified.
Another comment was about the controls being "mimicked". The commenter failed to mention the most important bit - that you can in fact use platform native controls if you wish. Some are inbuilt, and others are available as third-party open source components.
What this means is that you design a UI and then in platform-specific descendants - you can create variations of the one form or UI for various devices or classes of devices or platforms or whatever, it's quite neat - you can specialise to specify that on, say, iOS the edit box is a native edit box, on Windows it's a native edit, etc. Or, if the inbuilt controls (which are skinned to match the OS they're running on, yes) are good enough for you, you can leave it at the default and just use them.
Also, someone asked if Delphi is still being used - absolutely. I moved my consulting company over to using Delphi almost entirely a while ago, because demand was rising. Most of that seems to be in the cross-platform areas - most clients I see seem to really like writing an app once and then just varying the UI to make it correct per platform or device.