On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> Hi again... >> Strange ....how things work. MFC was a hottest technology in the market >> with the release of Visual Studio 6.0. MS has reaped huge unparalleled >> benefits from it. That was time and tide. Now the tide is with .NET. I admit >> that MS has done good in developing a good support for C# and VB.NET as >> complete languages. My only regret is that, in doing so, it has let down >> developers, who were once in plenty. I am talking about MS environment. >> There are a lot of C++ devs working on other environments. I only started >> programming with Visual C++ in 2005. I started with VS 6.0 and after a brief >> while switched to VS 2005. Back then MS showed some interests in enhancing >> the language further. It is not only about developing IDE, but about being >> compliant with ISO standards. As a person with a strong liking for Visual >> C++, I am sorry to say about the developments in C++ since 2005.
Hence why I still use VC++2k5 personally. I am thinking about upgrading to VC++2k10 though, if it has variadic template support though I will get it, if not, still debating... On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> I have even watched the videos on what is being done by Visual C++ team >> and their objectives. It seems that they are not in any mood to do something >> ambitious. For instance, let us talk about intellisense and code snippets >> features provided by IDE. The upcoming version of IDE has done one thing >> good that is removing the .ncb files and introducing SQL Server compact to >> buffer up the intellisense on the application load and calling only those >> features that are requested or close to being requested by the user. That >> would help the intellisense prevent breaking or searching for something >> every now and then. As the presenter in the video was saying about it and >> demoing it, it was waiting! Let us give him a fair chance and say that it >> will work. But why dunt they look at Visual Assist? MS gives a hell lot of >> reasons for autocompletion not working in VC++, one being C++ poor >> reflection capabilities. Visual Assist does that in style; I have not used >> it but I have seen it work on one of the MSDN videos itself. Also, MVPs >> working on VC++ advocate using Visual Assist. In order that I know about it, >> I asked and put up a question on MSDN dashboard and other forums with MS >> MVPs, but I did not get any answer. I simply cannot digest this discrepancy. >> MS itself uses C++ for many of their applications; earlier it used to admit >> it and now shying away from it. Introduction of tr1 library was just >> to divert the C++ coders from Boost regular expression library. Oh heck yes, I am completely addicted to Visual Assist. Until Eclipse or whatever gets something equivalent in functionality, it will *never* be useful to me. And I know many others like me who think the same way. On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> As I said, MFC was badly designed and MS did not do anything to improve >> it. MS kept reaping benefits till it continued to do so. Even today it is >> gaining them a lot, but it is certainly not overwhelming boom, I am not >> certain, but have read somewhere that Boost libraries outperform their STL >> counter parts. I have not read anything like that from MS C++ team since >> long. They consider it to be an achievement, if they catch up with >> something. .NET is a product of such a strategy. It started as a counter >> product for J2EE. In that process MS invented a new language C# quite close >> to C++ for C programmers and improved VB.NET taking it further from VB >> 6. Being a premier organization and having all the access to the most >> brilliant coders, it should have been pioneer in passing on the suggestions >> to C++ standards committee, but it is nowhere in the list. On the contrary, >> C++ wing in MS is sitting as if they were doomed and waiting for the >> standards to be out and then make a move. By then people look for >> alternatives. Yeah, they could have done so much recently, but Microsoft are not the pioneers that they were 15 years ago, in any area... On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> Coming to .NET ADOs and stuff, you are right. I did not know that we can >> work with ODBC in similar fashion as ADO.NET or can we? Can we have a >> disconnected architecture and caching capabilities? How performant is that >> using ODBC? You would need to build some of those functionalities on top of ODBC, but in general, ODBC is *slow*. It was not designed to be fast, and it has a limited feature set as it is designed to work with all databases (but none of them 'well'), that is why ADO.Net added so many features because it is trying to work around the slowdowns caused be ODBC. I would stay away from it. I had to work with it for a few years about a decade ago, it was horrible, and it has not changed... Wait until Wt::DBO comes out and adds more back-ends. :) On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> Good to hear that Wt supports XML. Although, I differ on the your opinion >> to not rating XML in good books as that is the only format I know when it >> comes to cross talk between 2 different databases over the web; binary is >> too much work. XML might be awfully slow. I suppose that you said that about >> XML it in context of presenting the data. If so, then yes I agree. XML has >> its own use in the form of tools like MS Build and Nant, code snippets, >> configuration files, etc. There are many representation formats like xsd, >> xslt, that help us understand complex designs easier. Ofcourse, this is my >> opinion with very little experience I have in XML. Not sure if build >> automation, configuration, data-porting, database schema design, site maps, >> etc. would have been easy enough without XML. Oh now I could write a book on just how horrible XML is. :) First of all, its beginning and ending matching tags are horrible for speed and usability for two reasons, one is because it increases data size, and two, because the ending tag is worthless. Honestly, what is the difference between this identical data structures in two different data format: XML: <my-tag attr=42>Hi!</my-tag> Sexp: (my-tag (attr 42) Hi!) There is never a need for end tags, and whoever thought them up in XML was retarded... Second, the tags add noise, just trying to read through them I can read the Sexp version easier, just because there is less syntactic noise. Third, the Sexp version parses a lot faster, it is a pure stack-based parser, and supports more construct then XML, like placement instead of just key names for 'attributes' and other such things. Now to compare it to protocol buffers. Let's say you want to save out a struct of data, like this: struct myStruct { std::string s; // "myString" int i; // 42 float f; // 3.14 }; In XML you might do this: <myStruct> <s>myString</s> <i>42</i> <f>3.14</f> </myStruct> Or more properly like this: <myStruct> <s><![CDATA[myString]]></s> <i>42</i> <f>3.14</f> </myStruct> In Sexp you might do this: (myStruct myString 42 3.14) Or if you wanted to be detailed in Sexp, although not everyone does it this way: (myStruct (s "myString") (i 42) (f 3.14) ) Or in protocol buffers in text mode: myStruct { s: "myString" i: 42 f: 3.14 } Or in binary mode (note easily representable obviously), it will only take up 29 bytes (mostly because of the string names, integers and such can be compressed, and the whole structure can be further compressed, it is designed to be very small, very fast, with as little overhead as possible). As well as the fact that protocol buffers has a few advantages over the other formats in that it handles all the parsing (no needing to iterate over your parts to fill in a myStruct) and is both backwards and forwards compatible as you add/remove/change parts of your myStruct. And I can keep going on for a great deal longer actually. :) On Sun, Dec 13, 2009 at 5:32 AM, Bhushan Inamdar <[email protected]> wrote: >> I am using IE 8.0 now. On reading your email, I checked out witty website >> in firefox 3.x and in that scrolling works well. I think it is being worked >> upon. Regarding the other problem, let it go. I will see when I work with >> latest witty sometime sooner... On Sun, Dec 13, 2009 at 9:37 AM, Ray Burkholder <[email protected]> wrote: > As a side note, only partially relevant to the original post,... if you are > looking for a performance based Windows GUI, and want to throw MFC into the > trash heap, you can also look into WTL. It isn't platform independent, but > is what MS uses on many of the tools supplied by the OS. WTL handles > multiple GUI threads better as well. > > > > Found at http://wtl.sourceforge.net/ along with http://www.codeproject.com > having many examples. Oh heck yes, WTL is so amazingly better. It was originally started by Microsoft, but they never continued it, however they released it, and people kept upgrading it, it is actually *fast* and with far fewer bugs then MFC. I use it at times when I am making something quick and not multi-platform (else I use wxWidgets, yes yes, I know, the bane of Qt programmers everywhere, and with an MFC'ish syntax, but it actually works far better then MFC still, and is native everywhere and with a license that I can *ACTUALLY*USE* unlike Qt's annoyance). On Sun, Dec 13, 2009 at 10:33 AM, Bhushan Inamdar <[email protected]> wrote: > Yes. I have had a look at it. Even downloaded it. But never really have a > chance to work with it. It is a wrapper around ATL as far as I know. But it > is much richer with its UI part. I had thought of working with it. But, > defered to MFC later due to its support. But since you have reminded me > about it again, I will relook into it. Thanks for that. Let me put in another chime for WTL, it is actually very well made, and it has great community support (not that I ever needed it, it actually works the first time when I do things unlike MFC, kind of a lot like Wt, Wt just seems to work). On Sun, Dec 13, 2009 at 10:33 AM, Bhushan Inamdar <[email protected]> wrote: > OvermindDl1, you said you came across a website that had a reference to > witty in its About section. Can you let me know any sites you know that are > build using witty? Has anyone else come across a website put to work using > witty? I have no clue what it was, I tried to remember the site, but I just saw it in passing once... The Wt website does need a "Users" section, it would increase the chance that most groups would use it if they see that it is in well use. ------------------------------------------------------------------------------ Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev _______________________________________________ witty-interest mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/witty-interest
