17 thoughts on “Currently dabbling with VS 2017.


  1. Unfortunately, the generics of FPC are useless to me. It’s missing some constraints, and I depend heavily on Attributes, so I can’t share Delphi source code with FPC – even if I’d want to.


    Besides, knowing more than language is a plus, and standard C++11 is the same everywhere.


  2. Roberto Della Pasqua Not as bad as it used to be before it was standardized. If it says C++11 or C++14 compliant – you can rely on that to behave as expected.


    The g++ compiler on Rasbian is C++11.


  3. Roberto Della Pasqua C++ has an ISO standard that everyone can count on. That means for most cases you don’t have to guess how and if some behavior is correct or not. Also there is a ton of people involved in developing new standards and it’s not just marketing driven as in “we just improve the language when it can solve some immediate problem we have” and thinking stuff through (I am not saying it is perfect but at least there is enough reflection over changes and benefits and possible side effects).


    Personally I find it refreshing to look into the C# drafts on github from time to time and read through comments and discussions because they show what things to consider in many regards and from different point of views. While there are also sometimes comments like “I don’t like this syntax, don’t do it it will lead to bad code blablabla” and the matching defending comments most of the stuff is really interesting and shows there is much to consider when adding things to a programming language and that in the end of the day it should help writing better code.


    Recent history has shown that while most of the things you can do with newer syntax could be done as well before but in a more complicated or verbose way. However many people have understood that writing less while not losing clarity is a good thing. And achieving that is not always easy but worthwhile.


    Unfortunately many Delphi developers rather wish for a component or control to save them writing code rather than for a better language itself that saves them from writing much code.


  4. Roberto Della Pasqua The ARM/ARM64 suggestion has been made many times, and not having Linux in Pro SKU, is incomprehensible. We can only hope Embarcadero change their mind. The VS 2017.3 Community edition is free, and supports Linux.


  5. For the record, I use the Enterprise version of VS as well as of RAD Studio. But – if you want grassroot support, you can’t be stingy with features for your free or low cost editions. A free Delphi that supports Raspberry would be a winner, even without DB support. On the other hand – DB support is a basic feature these days.


  6. Stefan Glienke “Personally I find it refreshing to look into the C# drafts on github”


    Python changes via “Python Enhancement Proposals”. There’s a mailing list to help your write one, and then you submit one to the development mailing list. Some of them read like published scientific papers with lots of references and footnotes. 🙂 An abstract of the idea, a background, a rationale, an explanation of proposed alternatives that were rejected, etc. Anyone can offer comments, join the debate, etc.


    There’s even a PEP for writing PEPs!


    python.org – PEP 1 — PEP Purpose and Guidelines


    And here’s one that added enums (at least the first that wasn’t rejected anyway):


    https://www.python.org/dev/peps/pep-0435/


    It’s amazing to be able to ask “Why is feature X like this and not like that?” and then actually go into the archive of PEPs and find a document explaining exactly why the feature was designed that way, a summary of the arguments for and against, alternatives that were rejected, etc.


    Someone once even gave a talk called something like “The Python That Almost Was” highlighting all sorts of PEPs that were ultimately rejected, including some that came close to adoption. Really neat.


    I have no idea how Delphi evolves today or who evolves it (sounds like the two project managers do it?).


  7. Roberto Della Pasqua The problem is that young coders don’t need to upgrade to a paid version of essentially any other language. Additionally, things like database connectivity and REST are already included in other languages; they’re not even going to want to look at something that doesn’t offer them. Kids today can start with a flash drive and empty hard drive and in short order they install a Linux distro and the compiler/interpreter/VM of any language you can name for free. Borland saw this coming a long way off.


    Jetbrains at least sells the IDE but the languages/compilers are free and even some of their IDEs are open sourced in community versions. By offering IDEs for many (popular) languages they can sell enough copies that the prices are very low and they can make a distinction between personal and corporate licenses. CodeGear seemed to entertain going that way (e.g. Delphi for PHP) but now EMBT is sticking with Delphi and C++ (when Microsoft completely dominates the Windows C++ IDE market). This keeps their prices extremely high which keeps new adoption down.


    It’s a real bad situation to be in. Honestly I think CodeGear had a better plan than EMBT and they wouldn’t have started a feud with RemObjects. Nick Hodges’ writings at the time suggested he wanted to modernize the language along the way RemObjects did while offering a compatibility mode for maintaining old Delphi projects.

Leave a Reply