Comments on The Spirit of Delphi: Breaking News: 1st November 2011 RAD Studio deal!

Isn't audits a sort of static metrics and does...

2012-02-08T21:05:42.524+02:00

Isn't audits a sort of static metrics and doesn't it contain most of what you're looking for? The problem with audits currently is, that it creates too many false positives where the issues are in the audits and not in your source code.

I've been in contact with Delphi from Delphi 1...

2011-02-13T17:45:50.068+02:00

I've been in contact with Delphi from Delphi 1, so I can say in short that Delphi 4 (or if you add optional parameters: Delphi 5), are the core language/IDE of Delphi 7. I'm not here to say that from Delphi 7 things get better, mostly D7 was likely the same that we perceive today XE: a fairly streamlined IDE, with a big package and was polishing the older versions. My point I wanted to make is that Delphi 2005 was considered a bad product for its bugs, slowness, dare to be VS like interface. I've been an Embarcadero developer (before CodeGear was bought) and I do know how they work internally (I've worked in the same office as I know that most of development is done, in Iasi), and in my view their direction of ditching bugs was a better way to do it. Is that nice to have many features but is that frustrating when you get tons of errors in development cycle. And Delphi 2006 was that kind of product.
Sadly (or better said: luckily) I don't use Delphi, I download at every 6 months Lazarus and I just read blogs about Delphi.
About Lazarus, I do think that is fairly hard to get the milestone you are proposing, because they have a small team (I've contributed a patch, was later removed, but I've been facing a lot of politics talk which was the killer switch for me to contribute more) and they work with a lot of runtimes: Win32, Qt, Gtk, Cocoa (if I remember right). The issue is that is really unlikely that they will get maturity without focusing on a specific features like: debugger, code completion, drag and drop done right, etc.
I do think that the price is high for Delphi, and I've bought for myself license for Visual Studio and I faced VS quirks too (WPF designer crash randomly in VS2008SP1, even was a feature that appear from 2007 as part of VS 2005, I mostly do WPF design in SharpDevelop beta, but seem to be fixed in VS 2010, yet my project was not migrated to .Net 4, so make it kinda ugly to work). Probably the price comes much worse when you don't compare Delphi with Studio, but you should compare the price of RAD Studio with VS which makes the gap even higher. Anyway, this at large I just think is an issue that can be addressed with multiple editions, the hard issue is to make a line for someone that do not want to buy a higher edition and wants that Starter to be just needed. In case you just create Forms like functionality and a database connection, nothing more fancier. VS have a very poor record of "customer support" in my experience, you will mostly Google things out, report bugs in Connect that will get in the next version. Also another problem in favor of Delphi, is that Delphi mostly do not need any runtime (kinda true with Vista/7 too, as updates push .Net 3.5 SP1). I've did face issues with incompatibilities across versions with VS.
The solution in my view for a Delphi developer that wants to push Embarcadero to do something? He/she have to ontribute to Lazarus, FPC or at least try to make sure that most of things you write will work inside Lazarus. This will make things clear for Embarcadero when an user will say: I don't need Delphi as gives too little for such a price. I would argue that Express Editions for VS come when Eclipse/Netbeans and Eclipse where on top of RIA/RCP and people were start switching. When Lazarus will be the tool that will solve bugs at least that to make it smooth to migrate most things to it, people will do it, and you may see later an Express Edition of Delphi. The red herring of Delphi is legacy code, so people will pay for it as they do for Cobol, if people did not rewrite everything in Java (or C#). If will be a full free implementation that migrates 99% assembly-free/platform agnostic code of your codebase.

Hello Ciprian, I agree with you about Delphi not b...

2011-02-13T09:29:14.459+02:00

Hello Ciprian,
I agree with you about Delphi not being intended to be sold at $150,-- but with the current functionality and support it should be sold at that price if you ask me.
About the Garbage Collector... I think this should be only optional, I truly believe that the developer is fully responsable of the life time of an object, but that's just me.
Lazarus is nowhere near the power of Delphi 7 IMHO, but if they try a bit harder they can "out gun" Delphi 2010 in 2 to 3 years if Embarcadero keeps coming with service pack releases i.e. XE version...
Since Delphi 7(yes it became a cliché), Delphi did NOT receive any real new features except for:
- Unicode(which isn't really that much of a deal...);
- generics -- came with a price, more bugs...;
- TBetterClass's in the IDE, search with a window explorer...;
- features that are practically stolen from third party tools for the IDE;
The above reasons pretty much covers my arguments regarding the price of the "Studio", if I were to go out and buy a IDE, I would have to chose between VS 2010 and Delphi XE, comparing functionality, "customer support" and price, I would be VERY inclined to buy VS, the only reason I would not is because I'm still optimistic about Delphi's future -- which currently is uncertain for me...
Regarding outsourcing, I have worked with a few companies from U.E. and U.S., and I can tell you that they would NOT buy licenses for developers, if developer(s) have license -- COOL, if they don't -- here's my license, work with it -- WAIT, wut?! yes...
Why don't they buy the license?! well... the price is very high... they could easily invest that money in bonuses, project, other tools WHICH give the best bang for the buck.
Something that really frustrates me the most, is the fact that Delphi has more developers in U.E. than in the U.S., yet we pay about 30% more for the product because if in the states is sold for let's say $1,200.-- and in U.E. is sold for €1,200.-- this is really annoying and a bad marketing attitude, Embarcadero clearly lacks:
- customer support;
- marketing;
- ambition to provide one of the best tools on the market, you can't charge a lot of money for a crappy and poor feature product, otherwise customers will eventually leave you... a very good example is the "economic crisis", in this period only the worthless companies went out of business, the companies with bad credit, bad customer support, etc., which do you think is NEXT?!

2nd part: my opinion about some things that really...

2011-02-12T17:31:44.606+02:00

2nd part: my opinion about some things that really I think that Delphi lost a bit the track compared with competition: Java and .Net.
- a simple framework to work multithreaded. An (optional) GC will help. I know that are people that disagree that GC is good, but is hard to do memory management with threads spawned: who have the ownership of the object? Is nice to find which thread leaves leaks at shutdown
- add a paradigm shift to make things to be cool around Delphi, I'm not here to say to add Linq like functional code or either to add eval support (to "JIT" Delphi code if is used with a DLL that isolates the compiler).
- add static hints to compiler (like Contracts/annotations that are checked at least at build step, if not live) and add basic features of Static Code Analysis. At large who buys Castalia, it's invited. I don't mean metrics, which are nice, but compiler based definitions as: "Cannot be null", "different from zero", etc. They mostly improve heavily old codebases quality.
Those are enterprise features that as far as I know are not in their Roadmap, but they do try to reduce the burden of developer or to make common mistakes.

I was living in Romania and I do think that at one...

2011-02-12T17:06:41.143+02:00

I was living in Romania and I do think that at one level Delphi was never intended as to be 150$ tool. Some reasons I found:
- vertical integration. If you buy an Apple device, you will get the same: a fairly high price but a tool that works from top to bottom.
- enterprise based tool: companies have money to sustain those products. Even in cases of outsourcing, companies will buy those licenses for offshore developers, so no lose at large
What I do agree with you, is that is non starter. Probably FreePascal may be the replacement (relatively :) ). I don't remember in highschool to talk about COM support, so for those cases FPC is enough (in the idea that you don't get all features that are too "sophisticated" for teaching). I'm not here to say that FPC/Lazarus is good, but is just enough.
About wasted money, most companies talk just about margins, and they do make a market study about how much can win and which is the price that market will pay. This is why Linux did not succeed on desktop, Windows or OS X did enough at their tagged price to "win" customers.
At the end, I don't necessarily think that Delphi necessarily need to have a low priced Delphi, as Lazarus exist, but to make more strides on competing the platforms on new platforms: native Android/iPhone development are a part of enterprise development, 64 bit, multiplatform (mostly OS X client and Linux server) and interoperability, but they are the next thing on roadmap.
One last thing I think Embarcadero did it in a really great job, is quality management. To take a 10+ years codebase and make it fairly decent and they don't rush for just features, is for me a great feat of them.