October 13, 2006
The Alternative Delphi Roadmap (by Simon Kissel)
Simon Kissel has published a very interesting proposal asking DevCo to invest in the Delphi Win32 (and Linux) more than in the .Net side. Very interesting appraoch.
First of all, I suggest you to read Simon's proposal, The alternative Delphi roadmap to success. (You can also see his original post in this thread).
What do I agree with? Well, most of it. I'm doing 40% win32, 40% Kylix (all my web stuff), 20% .Net (mostly Compact Framework), but I don't this reflects the average Delphi usrs. 90% of the companies I do training or consulting for are Win32 shops. Most of those 10% doing both Win32 and .Net use Delphi for the former and Visual Studio for the latter. And (with the exception of one company) they use only the ASP.NET part of .NET
What I do not agree with? I think we have to take more into the picture different types of development, in particular server side web development, which is growing at a very steady rate (AJAX, Ruby, PHP... not only ASP.NET). I consider .NET mostly a server side development system: is ASP.NET is much more relevant than WinForms, both for Visual Studio and for BDS developers. Is this going to change with Avalon (or WPF)? I doubt. My impression is that Avalon (now known as Windows Presentation Framework) is mostly targeting Flash and Web users, and in the words of Microsoft experts, it is not suited for the user interface of business/database oriented applications.
In other words, Delphi Win32 is still the best tool for client side (standalone utilities) and client/server (business data base apps). In this realm, Delphi for .Net and Kylix are much less relevant.
But if the world is moving to the web: has Delphi anything to offer? It might become an ASP.NET tool, but only if they can avoid following Microsoft at a distance, which is pointless. It might become a way to build PHP modules, as Simon suggests. I have my own idea (which will soon come out in another post). In any case, this is the area Delphi suffers and is loosing developers. Not in the user interface side (at least if you don't consider the CF, which seems still a minor platform).
15 Comments
The Alternative Delphi Roadmap (by Simon Kissel)
Simon is absolutely right. If I were Borland, I would change the roadmap based upon his advise! Trying to match Microsoft on .NET is a death sentense. As far as .NET is concerned, I would include ASP.NET 2.0 support and then add Mono support. Resurrect Kylix and focus 75% of your R&D resources on native code compilers for Win32, Win64 and Linux for Pascal, C++ and maybe even Java. Go NATIVE baby, go native. Get an Australian Aborigal and/or some aboriginal art for the next advertising campaign!Comment by on October 13, 05:59
The Alternative Delphi Roadmap (by Simon Kissel)
Read this link for a radical strategy for Borland, Win32 and .NET http://shortcircuited.wordpress.com/2006/10/08/borland-adapt-or-perish/Comment by M [http://thinkersroom.com/blog] on October 13, 11:48
The Alternative Delphi Roadmap (by Simon Kissel)
Eber: right now Borland failed to support .NET. MS easily outpaced them, because they control the platform totally. They already lost customers to VS anyway. BTW, it happend with VC++ and C++ Builder about fifteen years ago. MS started to deploy new features and Borland was unable to catch up in time. They didn't learn. To support .NET somehow, without the resources, using a "me too!" approach, and displeasing the current loyal base, they lost customers, and didn't get new ones. It's a lose-lose move to me. Kylix was a mistake because of the Borland approach. It was stupid to try to sell Kylix to Linux developers. The GCC/vi guy was beyond their reach (it won't use only Intel servers), and the LAMP guy not interested. The right move would have been to sell Kylix to Windows developers to port their server products to Intel/Linux - maybe bundling it with the Windows version, and not selling it as a separate product. I'm one of those who will buy a *working*, *fully interoperable* and *supported* Kylix. We already use a lot of Linux servers, and we'd like to port our server apps to Linux without rewriting them in C++.Comment by Kent Morwath on October 13, 13:05
The Alternative Delphi Roadmap (by Simon Kissel)
M: read my comment about that "radical" strategy. No one would buy a Delphi add-in for VS. Period.Comment by Kent Morwath on October 13, 13:06
The Alternative Delphi Roadmap (by Simon Kissel)
The post seems to suggest that DevCo can't do a better job of .NET than Microsoft. Aren't we using Delphi because we think they already make a better job of doing the native API than Microsoft? I think the abstraction of the VCL is the biggest advantage that Delphi offers. The future? Well, AFAIK the first Borland windowing framework was OWL. The framework allowed folks to migrate from DOS to Windows. Then Borland moved to Delphi and the VCL was born. Perhaps it's time for a new framework that's multi-platform and makes it easy to do WIN32, WIN64, .NET and Linux? There are other players here, but the success of Delphi was that one company controlled the framework, the language, and the development IDE. This allowed Borland to modify whichever element of the tool chain was most appropriate. This would still seem to be a unique position. However, is this a viable market? Most users don't need a multi-platform tool? However, I think there is a group of developers that would like to use a tool that doesn't lock them to a particular platform, even if they do end up only using a single platform. For example, I liked the idea that my tool of choice (Delphi) had an option to switch to Linux (Kylix) even though I haven't actually done any Linux development. So for me, Kylix added to the benefits of Delphi even though I hadn't used that option. Similarly, knowing that Delphi supports .NET is a good thing even though I still target native Windows. Realistically, I doubt I'd be interested in a tool that didn't target .NET. The downside is that a multi-platform tool might take a lowest common denominator approach that means it isn't good enough for any particular platform; obviously that's not going to work. So, is it a viable market? I really don't know how the size of that market can be determined, so I don't know. Some other quick notes. 1. Won't most users use the .NET platform to move to 64bit, rather than the native API? 2. I'm not yet targetting .NET yet but I've been giving serious consideration to moving to C#. Why? Well, for new projects, I can't see any benefit to Delphi. Does that mean I think Delphi should stop supporting .NET - no, absolutely no! However, it means that it needs to have a significant advantage. Is ECO the advantage? - I don't know but it has the potential. However, I do think it's inevitable that I'll move new projects to .NET so I'll be looking for a .NET tool. 3. Yep, I'm a non-upgrader; I'm still using Delphi 5. It simply does what I need it to do and some of my projects are still used on NT, so an older version works fine. Also, my users are internal and I don't need to keep up with the latest Windows look and feel. When I do, that's probably when I'll make the switch to .NET. It's also relevant that I'm also doing much less development than I used to and the upgrade costs became an approving problem. I like some of the new IDE features but there's nothing I can't live without and not enough to justify the cost of changing (including the hassles of components, etc).Comment by Bruce J Clark [] on October 13, 13:17
The Alternative Delphi Roadmap (by Simon Kissel)
Bruce: 64 bit early adopters will be those looking for raw power and large memory space. Those who see VMs, p-code (even jitted one) and garbage collectors as obstacles, not features. One day everybody will move to 64bits, even database frontends and web pages...Comment by Luigi D. Sandon on October 13, 15:14
The Alternative Delphi Roadmap (by Simon Kissel)
I sort of agree to Simon’s article even if I find the analysis a bit light I have to say that unfortunately this is true for Borland. I know it's true because I, like many others, was part of the people loving Borland tools because of the spirit, the great community, the excellent components (free wit source in many cases)... Now I to say that I was forced to live the Delphi world because my company switched to .net/C# What was my surprise to see that visual studio.net became an excellent IDE comparing to VS6 I kept doing Delphi coding on my free time and starting from that moment I started to dislike Borland products. And I hate them now. Why is that you may ask? I worked on big projects in visual studio (about 200 projects in my solution) and you could hardly make something crash! It was rock solid... Comparing to what was happening on the Delphi side: using Delphi 7 in a medium sided product using projects and packages I had to restart the IDE several time a day... Besides of all that, I could see big Delphi gurus leaving for C# or java because they could not stand the Borland tools any more. Big components vendor started to get out of business. The Borland newsgroups, Borland advocacy and TeamB started to get arrogant and didn't seem to care about people complaining about the quality of the products they loved. Borland itself got arrogant (the commercials at least). I can tell you a story about my contact in Borland. We switched to C#/VS but I was still interested in keeping an eye on Borland products. So I kindly asked my contact if he could keep me on the Delphi beta list and hand me some CDs when something came out... But no!!! He didn't want to... He (implicitly) said that because we where no longer paying customers he was not willing to loose some time with me... Even when I was his only chance for Borland to step in our company again some day. It is so hard to be part of the beta program. Borland had it all wrong. When you could just download the entire visual studio beta that you wanted, for free! No wonder that because of the small number of people that tried the product before it went final that it was that crappy and buggy... Ok, sorry, I could elaborate so much longer about how sorry I feel about what is happening to the Delphi product line... For me it's simply dead... What could bring me back to Delphi is certainly more Simon's roadmap then devco's What I want is "simple": *change your attitude, open your mind and listen to your customers *stop to run after what MS is doing, you always only come second like that *find your own way. If I want to do .net 2.0 or .net 3.0 stuff I will use visual studio (is there a chance you might wanna take a look at the prices or your products comparing to VS? What are the features you offer that are worth the price?) *remove passwords from everywhere and offer beta products for everybody. You should have noticed that your tentative to look your customers in your world is a failure *hire a real professional to draw your product icons and graphics. Admit that it's not a coder's work. Let him decide what the color palette should be. Ask him to take some time to redesign the various Borland sites, BDN and QA for example. Have you heard about a site called MSDN? Don't be ashamed to inspire from it (I won't tell any one). The point is that when I moved to the MS world I was so sorry for the people still in the Delphi world just because of the way the BDN site looked and let's not talk about its content. Most of the articles are poor (when you can call them articles) (sorry, I couldn’t resist) *stop adding countless "new" products doing the same things. Focus on one line and make it rock solid. Fast to code and even faster to debug (have you tried to debug a project heavily using interfaces at the same time then packages? I nearly felt asleep last time I tried to hit F7... *Stop adding language features to the Delphi language. They are more hacks then real OO... Ok, I would take the iterators in the win32 world *smaller updates, more often updates And now the concrete part, more objective... I only have two big wishes... *Give me a new kilyx version that is fully compatible with the current possibilities you find in the win32 part of delphi2006 (except native features of course). CrossKilyx should be supported...let me port my windows application to linux. I don't care building linux specific apps (maybe later) *about the .net part... All my projects are in Delphi win32 but I would like to start adding .net modules. So just let me use managed dll from a win32 host and also let me use a win32 dll from a managed host. Managed meaning Delphi.net but also C# and the like. Delphi can already do it why not but there is only one small sample showing how it works. Publish it, talk about it, document it, give us samples and give us some tools to help us on that (import/export managed/ unmanaged components). Also make this happen for kilyx. You heard about Chrome ? Hydra ? http://www.remobjects.com (Im not related to them, do I really need to say this ;) Good luck DevCo... ps: You will need more then luck...Comment by jonx on October 20, 16:06
The Alternative Delphi Roadmap (by Simon Kissel)
Jonk, I approved your post, even if with some doubts, as there are many statements I deem false: - "Restart the IDE several time a day" may be your experience, but not of everyone. I hear the same complaint for VS. Very subjective. - "Big Delphi gurus leaving for C# or java because they could not stand the Borland tools any more." Who? They moved because they had good job opportunities, many with Microsoft, not because they "hate" the tools. I know many of them, I don't think you are interpreting them correctly. - "Big components vendor started to get out of business." Also many .NET component vendors have problems, and some who moved from VCL to .NET later switched back... or kept devoting time to the VCL again as it is still a good revenue stream. - "It is so hard to be part of the beta program." Not everyone has Microsoft resources. Is this looking for betas of for free products? Get the Turbos! - "Stop adding language features to the Delphi language. They are more hacks then real OO". Should they leave all new feature to C#? If private-that-works, operators overloading, class data are not OOP features, what are they? I have to say I appreciate and agree with some of the other points (Kylix, not running after Microsoft...). Also, I agree on mixing managed and non-managed code. Do you know Delphi has "managed exports" built into the language? But you are right, should be a more prominent feature!Comment by Marco Cantù [http://www.marcocantu.com] on October 21, 01:30
The Alternative Delphi Roadmap (by Simon Kissel)
The logic of Delphi was that it depended on nothing other than the product, as experienced by most users. The primary logic of using Delphi is for the users. It was/is a way of communicating that allowed users to experience an application, allowing the developer full control of the user experience without a hidden layer of the OS. In a way it was rebellion, that then led into Kylix, maybe a flawed business model, and less effective on Linux than Windows(Delphi). Simon Kissel is correct in his analysis not just for the 'Roadmap' (economic reasons) but also because Delphi was an intellectually profoundly independent product. It gave (gives) developers freedom to deliver/deploy apps that are unique in that they give the user the experience that the developer intended with no overhead. Don't forget the user. The beauty of Delphi is that the user believes that the app provides him/her with freedom. dotNet does not do necessarily do this for an average user. As with many companies Borland is loosing its focus on the 'bottom end', arguably the most important 'end' of its market.Comment by Angus Jack [http://www.bre.co.uk] on October 25, 16:53
The Alternative Delphi Roadmap (by Simon Kissel)
I need fast and native RAD Delphi Win32/Win64 IDE with good unicode supprot. I do not need any ".NET" crap. I also need only command line Delphi compiler for Linux, but I can use free pascal for this target. So Borland please focus on native Win32/64 and throw away ".NET" and Linux.Comment by Simon Bodrov on October 27, 11:18
The Alternative Delphi Roadmap by Simon Kissel
Here's a really stupid idea - have the CodeGear/Borland team sponsor freepascal to do bsd/linux development and stop fighting - keep fpc as a separate compiler and make money through Borland/Codegear consulting for Linux. Keep sales of the windows IDE going.. but try and make friends and not fight on the linux front. Flamebait? Possibly. Study IBM and linux though.Comment by L505 [http://z505.com] on July 25, 10:47
The Alternative Delphi Roadmap by Simon Kissel
I love delphi, but I need to do program in other platform too. I will keep Delphi for my Microsoft customer ( for now ) as one of my IDE beside the origin VS. I will do Linux on someting else. There is a lot going on in this area too and I think it's wrong to blame Delphi not to compete in this area. It's their wright to be just 'Another Microsoft Tools'. Beside there is no standard gui / not as good as Microsoft did in windows. So, the move will be so difficult and money / time consuming. Face it. It's just impossible. But market is market, and my customer don't care about that, they just want linux, so I have to made software on something else ( realbasic or maybe RIA ) I think Codegear should not put the Motto 'A new organization 100% dedicated to the Developer Tools Market'. It's just doesnt fit, maybe put a word Microsoft will do. : 'A new organization 100% dedicated to the Microsoft Developer Tools Market.' Hmmm ...yes that's wright. sweet dream.Comment by dreamer on October 17, 06:26
The Alternative Delphi Roadmap by Simon Kissel
Oh, forgot something. Read Delphi road map has 'compile for another os platform' ( afer commodore, last thing they will do ? ). Should we wait for theme ....? Are they serious ? Hearing from what they have said in various discussion.... hmm I don't think so. CG should remove it from the road map if they don't mean it. please don't make false promises to people.! let them go guys, let them go. they also need to make a living. sweet dream.Comment by dreamer on October 17, 06:40
The Alternative Delphi Roadmap by Simon Kissel
I would have to say, that whatever parts of the proposal are right or wrong, I would have to say that he hit the nail on the head almost all of the way through it. I say that because as I'm reading it I'm thinking, yeah, that is what I want. I mean really, price of Kylix was bad, didn't want to author code under a Linux GUI, and Borland always seemed to be late to the table with the features that would make me buy. Really to me there are largely two "factions" in the Delphi software development world (that I have seen). There are those that develop software for internal use, and there are those that use it for products they sell. The internal users have less of a demand for Linux desktop apps and more of a use for server-side programs. The software vendors that develop generic vertical or horizontal applications are more the ones that want cross-compilation to other platforms. That is natural as it immediately increases your audience. RealBasic is a good example of what I want. From a main menu of the application you simple select to compile the application for Windows/Mac/Linux. Does nobody at CodeGear take notice that FPC/CrossKylix/Lazarus have grown so much. I mean, really, keep in mind that the group of people that will explore these options instead of just going over to VS.Net or Java are devoted Pascal lovers. CodeGear Execs can say what they like. History is being written, and the consensus is that Borland/Inprise/CodeGear has been dropping the ball consistently since the Delphi 5-6 era. That not just my opinion, I hear that in nearly every conversation I have that approaches the Delphi/Borland subject. Its always comments about how great the tools are/were and how badly they were marketed/supported. It makes me soo frustrated that the tools are awesome but the company is single handedly destroying a great legacy. Its like they just have closed meetings and discuss direction without having a good understanding of the users. Marco, I don't know how you feel about it, but its my opinion that CodeGear should take some SERIOUS time to rethink their angle. I would say you along with Simon Kissel, maybe Bob Swart, Zarko Gajic, and a few others should be consulted on a good direction for CodeGear and Delphi in general. Server-side, I want cross platform with database connections to MySQL/Postgres, and ability to natively compile Apache modules. On the desktop side of things, I want to be able to compile for as many platforms as they can fit under the hood, including .Net, and I want to do it with a unified VCL environment. Thats it. Right now I work in a $7B Fortune 1000 company that has been forced through the various "We're selling our code tools" announcement to migrate everything in Delphi to Java within 2 years. Execs, Directors, and PMs all the way up the chain want to dump CodeGear tools because of massive fear of a soon to be discontinued product. Surely someone in CodeGear is willing to listen to something other than the sound of their own voices? Marco, I'm interested to hear what your perspective is on why CodeGear is on such a downhill slide? Thanks for your many years of great (and the best) books on Mastering Delphi, they have truly been my tutor for all-things Delphi.Comment by Angus McGinnis [http://www.surevital.com] on January 11, 02:54
Post Your Comment
Click here for posting your feedback to this blog.
There are currently 0 pending (unapproved) messages.
The Alternative Delphi Roadmap (by Simon Kissel)
Comment by Eber Irigoyen [http://ebersys.blogspot.com] on October 13, 03:15