Future of Delphi, VCL.Net and WxF

Monday, 25 June 2007 13:30 by salim

I have been reading a lot of articles about the recently published Delphi Roadmap. It was interesting to see so many people happy to know there is an actual Roadmap for Delphi.

Codegear(Borland) has been extremely tardy and unprofessional during the last few releases of Delphi. Delphi 8 was a total disaster, D2005 was only slightly better. D2006 did make things a bit better and D2007, as far as I have seen seems to be at least as stable as D2006. Publishing the roadmap was a big sigh of release for me who has to constantly address the apprehensions of product owners about the future of Delphi.

In these discussions, this from Marco Cantu helps us reading the tea leaves better.

Since I never used Delphi Winform support, the demise of it was not a big item for me. I, like Marco believes that Delphi is VCL.

While there is a flux in UI paradigms in the .Net world, there is a very significant overall direction. The close connection between WPF, XBAP, SilverLight etc. makes it possible (more than any other time) to create common UI elements for Client Side and Browser hosted applications. Considering the push from pretty much everywhere to a more service oriented application architecture, this is a significant win.

At the current state of affairs at CG, it looks like Delphi and VCL.Net will be atleast one release cycle behind the latest from MS. Making VCL.Net platform neutral, atleast within .Net, Win32, Win64 circle is crucial to support fast changing technology outside the control of CG.

I am not sure this is what I see. For e.g .Net compatibiltiy breaking changes introduced in D2006 Win32 and beyond for Unicode, like WideStrings.

Ideal VCL will be write once, compile in win32/64/.net/Mono. With accompanying interoparability solutions, I can use Delphi as the only tool I need to provide most solutions, provided the IDE will stop crashing so much...

Well, I guess we will see in the coming months, both from Microsoft and CG.

Actions:   E-mail | del.icio.us | Permalink | Comments (0) | Comment RSSRSS comment feed

Sextuples... A Miracle?

Wednesday, 13 June 2007 01:23 by salim

It is a well known fact that there has been an increasing trend in multiple child pregnancies including several Sextuples in the last several years. Today there was another one reported in the morning show. I found it rather disturbing to see the dilemma of that couple. Of course, I guess they are not allowed to express any major concerns in morning television.

Every time I hear about these extreme multiple pregnancies, which almost always happens in relation to fertility clinics and artificial insemination, I wonder why it is not considered as a failure of service? I understand the emotional requirement to have one child or two, but it is hard to imagine that an average family will be overjoyed by having six babies at one go.

I know it cost a lot for one impregnation at the clinic, but selecting a potentially dangerous decision of allowing too many embryos implanted is appalling. From both the clinic's side and the couples. The danger is not just for the parties involved, but the babies as well.

Such incidents should not be considered a miracle or abundance, it is yet another failure of the healthcare system. Here is a very interesting article about this issue.

Actions:   E-mail | del.icio.us | Permalink | Comments (1) | Comment RSSRSS comment feed

It is a beta, look how cute it is? Please don't hurt it...

Wednesday, 13 June 2007 01:09 by salim

Today several bloggers commented about bugs in Safari Windows Version Beta.A controversial one by David Maynor and Inaequitas, many news postings and a whole bunch of comments. One argument, that was repeated in a lot of places, as a basic ethical stand when dealing with security issues, is that, one should not be divulging security and other issues to the public during beta/ctp.

I have been participating in pre-release programs with and without NDA. In all projects that I have participated, there were never an ethical issue for posting the findings in public forums. In most cases it was considered healthy. I remember other incidents like several Symantec reports about Windows Vista security issues very early in the CTP stage.

In my opinion, when a company publish content to the public space, whether it is labeled as alpha or beta or CTP or release without explicit binding constraints, it is in public domain. I do partly agree with David that providing information about issues with software in public domain should not be considered as a bad practice. If the company that produced the software is incapable of responding to it in a timely manner, they should be held responsible.

I always wonder, what other industry in the world will have a statement like "this product is sold as-is, ...." and a fool proof indemnity clause as a standard statement of service.

If software industry want to be as reliable and predictable as other industries (like aerospace for e.g), it need to shed the special status it always crave. It is just another industry. I work in a factory, that looks very nice, makes me happy and gives me good amount of money. I have been to enough discussions that elevates my vocation into a great intellectual exercise. Now I want to produce. I do not have much sympathy towards other software vendors. If you cannot survive in a demystified market place, you will simply wither away.

Actions:   E-mail | del.icio.us | Permalink | Comments (0) | Comment RSSRSS comment feed

At last, Highlander (Delphi.Net) Roadmap is out

Friday, 8 June 2007 21:15 by salim

For quite a few months of conspicuous silence about the plans, features and roadmap about their upcoming releases (even a few months ahead), CodeGear finally have published their roadmap for the forthcoming releases.

See Nick Hodges blog here about the publication.

First of all, I am very relieved to see it published. Without a straightforward answer or published roadmap, I was having hard time planning for technology adoption for our oncoming releases. There was a steady stream of complaints and sometimes angry venting at the Codegear forums about the disappearance of the roadmap.

Well, now that it is hear, it is time to start more educated speculations about it :-)

I am specifically interested in the Highlander release roadmap, which is supposed to come out by the end of this year. The farther releases, understandably has much less clear specs, except of course the promise of 64 bit.

Here are some key points I noticed

  1. DBXClient, this is a very interesting feature. It is nice to have a thin driver layer that is native in both win32 and .net. Unicode support is great as well.
  2. Consumption of generic types - This is nice, but does it mean there wont be language support for declaring generics?
  3. CodeBehind support for Asp.Net, very nice
  4. Aspx provider model for fod DBX4 seems promising.
  5. I am not sure what unified metadata, but if this further reduce the incompatibility between ado.net and DBX I am happy.
  6. Is SQLDatabase the replacement for Interbase? It is apparently posed as an alternative to SQL Compact edition from MS. I have to see how that spans out.
  7. One thing that is not clear is how these features span among the SKUs. I will be very sad to see if the DBXClient and SQLDatabase is limited to the higher SKUs.
  8. I have never been a big fan of ECO, mainly because of its strong dependencies and issues incorporating it into TDD. However, VCL.Net support for ECO is interesting. I am sure a lot of people who use ECO will like that.

I overall satisfied with this. Lot of the technologies that I am planning to move into like SmartClient Software Factory, Workflow Foundation and WPF (I am sure there wont be a WPF design surface available, but that is not a major issue for the next one year for me) will work with Highlander.

Now we have to see how true the release time line will be.

Actions:   E-mail | del.icio.us | Permalink | Comments (0) | Comment RSSRSS comment feed

And we are back to old rocket

Thursday, 7 June 2007 21:42 by salim

I have always wondered the logic of space shuttles. That thing cost almost $2 B (not counting design and development cost) to build and about $500 M to launch. Most of the stuff that is used during the launch is not coming back either. At the same time Ariane cost about $600M per build.

And then we have the "next generation" space exploration vehicle, which will be (with three quarters of probability -sic- will be completed in 2015) which, of course will cost about $15 B to build.

All these should be compared to the $18M X-prize winning spaceship. Yes, yes I know, that was experimental, went only to the outskirts of earth, technically, outer space. So, how come when NASA build something, it becomes a monstrosity?

Then I got a few tidbits dating from Nixon era, how he wanted to build something that is specifically suited for secret spying satellites and various classified military missions. That makes everything so clear. If you want to make something that works 100 times less efficiently with 1000 times more money, all you need is a classified project funded by American military and NASA !!!! 

Categories:   Science | Space
Actions:   E-mail | del.icio.us | Permalink | Comments (0) | Comment RSSRSS comment feed

Evil Evil Microsoft...

Friday, 1 June 2007 22:45 by salim

In Martin Fowlers blog there was this article about Ruby and Microsoft. While pondering the possibility of Microsoft implementing a .Net implementation of the Ruby Platform. There is this one sentence that caught my eye. Here it is

[snip]..but Microsoft imposes drastic limitations on its employees' ability to download open source software, let alone look at the source.[snip]

He makes this observation while talking about the difficulty in porting the ruby runtime. I am wondering, can someone do this these days? Put a ban on all its workers from browsing to open source projects? I am sure they have a very complicated, comprehensive filters around their boundaries. There are also reports of dogs and cats getting into MS campus, never to be heard from again!!!

Actions:   E-mail | del.icio.us | Permalink | Comments (0) | Comment RSSRSS comment feed