BCB3: First Impressions
Peter Steele
Well, I can only say that I'm very very very disappointed. This feels more
like a minor version 1.5 upgrade instead of a full version upgrade. Is there
any significant new feature in this release that would warrant the cost to
upgrade? I've discovered nothing that will increase my productivity.
There is *still* no source browser,
there is *still* no "step out" debugging command,
there is *still* no "add new class" command,
there is *still* no windows menu (the unsorted tabs on the main edit window
are a royal pain),
there is *still* no real support for directories for organizing
projects/packages,
the text editor is *still* not as good as most of the competition,
and this time around we got no printed documentation for the VCL.
I do like the concept of packages, but I'm not sure if project groups have
much use (I'd prefer just to be able to open multiple projects like in
JBuilder). And it's nice to see drag and drop text in the editor and a "Find
in Files" command. There are so many other things I was expecting though in
this update and all I can say is that I'm quite disappointed. I don't really
see anything to warrant even keeping the upgrade, especially considering the
cost and especially considering we'll have to fork out more money to get the
printed docs.
Granted, I've only been using the product for a few days so I'll stick at it
for now. Perhaps I'm yet to discover some new important features that'll
make me change my mind about this product. We'll see...
Chris Uzdavinis
I have a sharply different perspective. The upgrade is major and I am very
satisfied with it. What version do you have? If it's "standard" you
probably are not seeing all the improvements that have been made. I have
the pro version and like it a lot.
* Stability.
It is so much more stable that I can't compare it to 1.0 at all. I have
never crashed 3.0 yet, and 1.0 died all the time. By itself, this is
enough to increase productivity.
* Speed
The debugger no longer lags when you have watches set. This saves time,
and also increases productivity. ToolTip evaluation is nice and
convenient... another time saver. The local variable inspector reduces the
need for me to set watches, and is easy to use. All this put together
saves me probably 10-15% of my time debugging.
* Compatibility:
The language more closely follows that C++ standard, supporting new
features and a totally revamped standard library/STL.
* Bug fixes
Lots of bugs have been fixed. This reduces the requirement for workarounds
and independent discovery of workarounds.
* Documentation
Better, but still has a way to go.
* Components
The internet components are better and are native VCL
* TASM
Inline assembly is now a possibility for me.
Hmmmmm, some of the other features (like ActiveX stuff) are probably nice
for other people but I don't use them. What I listed above are what I
found most beneficial to the upgrade, and I'm very happy with it. I also
don't use the database aspect of BCB at all.
Chris
Peter Steele
>What version do you have
I have the Pro version...
>* Stability.
>It is so much more stable that I can't compare it to 1.0 at all. I have
>never crashed 3.0 yet, and 1.0 died all the time.
I crashed it twice this morning so far, trying to port my old version 1
components. I crashed 1.0 a few times as well, but it was never what I would
consider unstable, at least in the work I was doing.
>* Speed
>The debugger no longer lags when you have watches set.
I haven't really had a chance to test this. I currently trying to get my old
1.0 app to work under 3.0. I have a bunch of components that need to be
ported to the new package concept and one component is having linking
problems. So I haven't had much opportunity to even see how the debugger
works. Like I said though, my comments were "first impressions"...
>* Documentation
>Better, but still has a way to go.
Why of why did they decide not to include hard copy?
My complaints largely are aimed at the IDE, which I felt in 1.0 was very
very weak compared to the competition and had expected a major improvement
in this area, especially when I read all the promo material. It sounded like
a totally new project manager/editor. There is virtually no change at all
though to these two key components of the IDE and I will probably still have
to use a third party tool for some of my work. I have never felt the need to
use a third party editor when using Visual C++ or Visual Basic. JBuilder is
also much better. C++ Builder is not a state-of-the-art IDE, in my opinion.
Mind you, I much prefer using C++ Builder instead of Visual C++ and MFC.
Even with its weaker IDE BCB is much more productive that VC++. Give BCB an
improved IDE and it would truly be an outstanding tool....
John Croudy
Peter Steele wrote:
> I have never felt the need to use a third party editor when using Visual C++
> or Visual Basic.
I just thought I'd mention that I think the BCB editor is far better than Visual
C++ editor. For one thing, it can do *quick* record/replay and *column* cut and
paste. I've saved so much time since I discovered that.The 'open source/header
file' always works, and I love it. In VC, there was some kind of 'open header'
and it never worked properly, unless your source file was part of the
ClassWizard database or something stupid like that. I never understood what
classes had to do with finding the right header file. The tabs at the top of the
window are a great idea. I hated having to go to the Window menu in VC. Then
again, using Ctrl/Tab to cycle among files is great in both BCB and VC. The only
things I don't like about BCB's editor are the way the cursor doesn't wrap
around at the end of the line, and the fact that it doesn't have a splitter bar.
As for the BCB (1.0) Ide generally, it leaves VC++ far behind. It actually
*does* all the grunt work that VC tries and fails to do with it's futile
ClassWizard, WizardBar and ClassView nonsense. WizardBar... it slows down the
text editor unbearably, even when 'disabled' and it helps you find your
functions! So does the 'Find' dialog. I can find my way around my code without
some slow clunky robot to 'help' me. In VC, most of my attempts to use ClassView
resulted in error messages because some stupid file was out of sync with some
other stupid file. In BCB, the object inspector works and it works properly.
In my opinion, the weaknesses of BCB are it's debugger, the diabolical
documentation, and some of the weirder aspects of VCL that cause strange crash
bugs until you figure out a workaround. The BCB debugger is a nice attempt but
doesn't 'cut it' I'm afraid. Here, VC++ shines. By the sound of it, there's no
point in me getting BCB3 because the things that I don't like about BCB are
still there. I want a decent debugger firstly, and some decent docs. Failing
that, I'll stick with version 1.0. The devil you know is better than the devil
you don't know...
--
John Croudy <john....@semitechturku.cod> 'cod' should really be 'com'.
http://www.freepages.cban.com/pinkmouse/
The views expressed here are my own personal views and do not in any way
reflect the views of Semi-Tech (Turku) Oy.
Exception
John Croudy <john....@semitechturku.cod> wrote:
* I just thought I'd mention that I think the BCB editor is far better than Visual
* C++ editor. [...]
Despite the advantages you mention, the BCB code editor has an enormous drawback (in
my opinion), and that is the lack of being able to use proportional fonts!
Which Visual C++ does, by the way.
As a developer I have to spend many hours looking at source code, and those ugly
clunky monospaced fonts are an insult for my aestetic appeal.
Borlanders, are you listening? I want to use proportional fonts in the code editor!
I'm willing to give up column cut/paste for that. (Make it an option).
Please, please, please.
---
Design your Exception class hierarchy well...
VCL+MFC+API+OCX == new *Modern_Developer;
Neil Booth
John Croudy <john....@semitechturku.cod> wrote:
>In my opinion, the weaknesses of BCB are it's debugger
Have you seen the debugger in BCB 3? It's streets ahead of that in BCB
1.
Neil.
Peter Steele
>The tabs at the top of the
>window are a great idea. I hated having to go to the Window menu in VC.
I would be much happier with the tabs if they were sorted. As it is, their
order is essentially random and when I have a lot of files open and some
tabs are hidden I don't know if I should scroll to the left or scroll to the
right to find the file I want. I usually end up going to the list of files
in the project manager and clicking on the file there since it is quicker
than trying to find it on the tabs. Unfortunately, it now looks like the
files in the project manager are no longer sorted... :-(
What I'd really like is a quick popup menu or pulldown menu right on the
editor window with a list of open files in it, and another pulldown menu
that lists the methods of the current file. Visual Basic has had this
feature for a *long* time, as did LightSpeed C on my Mac back in 1986...
Jody Hagins
Peter Steele <pe...@ats-forest.com> wrote in article
<6em4ad$qi...@forums.borland.com>...
[SNIP]
> What I'd really like is a quick popup menu or pulldown menu right on the
> editor window with a list of open files in it, and another pulldown menu
> that lists the methods of the current file.
That's what I want, some popup with a sorted list of methods. Under vi, I
always made sure that functions began at the first column, so I could do a
quick, /^whatever and I'd go right there. Even with the extra typing it
was OK, because it was all at my fingertips. I can't come close to that in
the BCB editor. Also, I want multiple copy/cut/paste buffers, like vi, so
I can copy several things at once from one file, switch to a few others and
paste them from one of my 26 available buffers.
Kent Reisdorph (TeamB)
>>Despite the advantages you mention, the BCB code editor has an enormous
drawback (in
my opinion), and that is the lack of being able to use proportional fonts!
Huh? How could you possibly write code with in a proportional font? I'd hate
to have to read it.
--
Kent (TeamB and TurboPower Software)
-----------------------------------------------------
BCB Solutions Web Page - http://www.turbopower.com/bcb
Alex Bakaev [TeamB]
What features editor is missing ? What IDE has those editor features ?
Alex
Peter Steele wrote:
[snip]
Alex Bakaev [TeamB]
Chris Uzdavinis wrote:
>
> The one thing that VC++ has that I still miss is the ability to put a
> breakpoint on memory, and have the program stop when that memory chages,
> regardless of where the program is in your code (or the system code).
I was under impression you can put memory breakpoints in BCB 3 :)
Alex
Alex Bakaev [TeamB]
John Croudy wrote:
>
> In my opinion, the weaknesses of BCB are it's debugger,
That might have been true of BCB 1.0. Have you tried BCB 3 debugger ?
Alex
Chris Uzdavinis
> my opinion), and that is the lack of being able to use proportional
fonts!
Proportional fonts? For programs? Whoa, I thought I had heard it all!
Personally, I can't stand proportional fonts with code because nothing
lines up correctly and everything is difficult to read.
To each his own, I guess. <shrug>
Chris
Chris Uzdavinis
> In my opinion, the weaknesses of BCB are it's debugger, the diabolical
> documentation, and some of the weirder aspects of VCL that cause strange
crash
> bugs until you figure out a workaround. The BCB debugger is a nice
attempt but
> doesn't 'cut it' I'm afraid. Here, VC++ shines. By the sound of it,
there's no
> point in me getting BCB3 because the things that I don't like about BCB
are
> still there. I want a decent debugger firstly, and some decent docs.
Failing
> that, I'll stick with version 1.0. The devil you know is better than the
devil
> you don't know...
I agree that the weaknesses of BCB 1.0 is is debugger. But BCB 3's
debugger is quite a lot better. VC++ still has a slight edge, but BCB 3
almost makes up in every way.
The one thing that VC++ has that I still miss is the ability to put a
breakpoint on memory, and have the program stop when that memory chages,
regardless of where the program is in your code (or the system code).
But BCB 3.0's tooltip evaluation is better than VC++ in my opinion in
several ways. I didn't notice until last night, but evaluation of classes
shows all the members, not just a "..." like VC++ does.
Also, it works better with members and properties.
for example, if I put the cursor over the "x" in "foo.x" it gives me the
member x, not the local variable x. In VC++ you must highlight foo.x to
get the expression evaluated.
It is REAL nice. The debugger is quite noticably faster too. And more
robust. I haven't been able to crash it even once since I got BCB 3.0.
Chris
Chris Uzdavinis
Really? I totally overlooked it. In that case, I don't know of any reason
that VC++ has a better debugger then BCB 3. That's really cool! Memory
breakpoints were the one thing that gave VC++ the edge in my opinion, and
apparently now it's gone. Excellent!!!
Chris
This is fantastic. I'm very happy now! I had no idea that was there.
It's fairly hidden though, which explains why I missed it. From Run | Add
Breakpoint it is not available and there is no hint of it. The only way
I've found the ability to set data breakpoints is to right click on the
breakpoint-view window and add an address breakpoint.
Whew, I'd have never known to look for it since it was not obvious to find.
Had you not mentioned it was there I'd have NEVER found it. Thanks again!
Why such a feature is hidden is strange, but at least it's there!
Alex Bakaev [TeamB] <al...@jetsuite.com> wrote in article
<350EE9CC...@jetsuite.com>...
> Chris Uzdavinis wrote:
> >
> > The one thing that VC++ has that I still miss is the ability to put a
> > breakpoint on memory, and have the program stop when that memory
chages,
> > regardless of where the program is in your code (or the system code).
>
Exception
"Chris Uzdavinis" <chr...@uzdavinis.com> wrote:
* > my opinion), and that is the lack of being able to use proportional
* fonts!
* > Which Visual C++ does, by the way.
*
* Proportional fonts? For programs? Whoa, I thought I had heard it all!
* Personally, I can't stand proportional fonts with code because nothing
* lines up correctly and everything is difficult to read.
Well, if my programs would have to rely on 'lining up columns of code' for
readability, I'd retire! Right now!
It's not FORTRAN or COBOL anymore, where positional syntax was mandatory... These are
the times of the well-chosen variable names, who act like actors in the play of your
application, this is the era of classes which materialize your structure, these are
the times of the Design Patterns which form the plot of your program, as if it was an
unraveling novel!
Oldfashioned monospaced fonts are something of the past, at least for me.
* To each his own, I guess. <shrug>
That, I agree with you heartily.
Exception
"Kent Reisdorph (TeamB)" <ke...@turbopower.com> wrote:
* >>Despite the advantages you mention, the BCB code editor has an enormous
* drawback (in
* my opinion), and that is the lack of being able to use proportional fonts!
*
* Huh? How could you possibly write code with in a proportional font? I'd hate
* to have to read it.
Why? Humans are used to reading proportional fonts. Fonts with serifs on them will
even read better: your eye is guided from one letter to another. It's not without
reason that books are printed in proportional fonts. And since my PC can do font
smoothing, I enjoy my screen which looks typeset instead of typewritten.
In my career I've met a lot of people who didn't understand my desire for a
proportional font. Their stance: "You can't line up columns of program code that
way!". Well, I work with objects, not 'columns of code'.
A good program reads like a novel, and good type enhances the experience.
Ah well... There's no good in discussing preferences. As the Dutch say: "Over smaak
valt niet te twisten", the Germans say: "Ueber geschmack laBt sich nicht reden", or
in English: "Beauty is in the eye of the beholder".
Chris Uzdavinis
> Well, if my programs would have to rely on 'lining up columns of code'
for
> readability, I'd retire! Right now!
I didn't mean for Fortran-like code :) I mean I like it to line up when I
indent with 2 spaces, the blocks line up. Or maybe this:
int i
unsigned long x; // this example comment is long and therefore
// spans 2 lines and I want the comments to be
aligned.
I'm typing this with proportional fonts, and I want the "i" and "x" to be
aligned. Can't quite get it. Maybe it's REALLY messed up for some
people's displays above. The i is approximately above the x on my screen.
Same with the comments. I want the 2nd row to be directly below the 1st
row of comments. That will move around on different displays as well.
> Oldfashioned monospaced fonts are something of the past, at least for me.
They are cutting-edge technology for me. :)
> * To each his own, I guess. <shrug>
> That, I agree with you heartily.
I use Emacs and some co-workers use vi. I learned tolerance a long time
ago through necessity. <g>
Chris
Chris Uzdavinis
> Ah well... There's no good in discussing preferences. As the Dutch say:
"Over smaak
> valt niet te twisten", the Germans say: "Ueber geschmack laBt sich nicht
reden", or
> in English: "Beauty is in the eye of the beholder".
Hmmm, that's funny. I've always heard the expression go as, "Beauty is in
the eye of the person looking at properly aligned columns."
<grin>
Chris
Kent Reisdorph (TeamB)
>>It's not without reason that books are printed in proportional fonts.
Yeah, but have you ever seen one in which the code examples were in a
proportional font? ;)
Jody Hagins
Exception <Exce...@mailcity.com> wrote in article
<350e40e2...@forums.borland.com>...
> John Croudy <john....@semitechturku.cod> wrote:
>
> * I just thought I'd mention that I think the BCB editor is far better
than Visual
> * C++ editor. [...]
>
Since we are talking about the editor, I *hate* clicking on some white
space at then end of the line, and it positioning there. If I click on
some of the white space area at the end of a line, I want the cursor to go
the the last now-whitespace character.
As of yet, I've not found that option... (OK, I click, then press 'End'
but I want it to do this without the extra keypress of 'End')
Ralph Friedman (TeamB)
Jody:
...and then what do I do what I want to click on some white space at the and
of the line and type there? Pressing end w/o the extraneous click will do it
for you a bit more efficiently.
--
Regards
Ralph (TeamB)
Herrsching, Germany
Jody Hagins wrote in message <01bd520e$b53aeec0$39ca06d0@fpfzqlga>...
John Croudy
Jody Hagins wrote:
> If I click on some of the white space area at the end of a line, I want the
> cursor to go
> the the last now-whitespace character.
Yes, me too. It should be an option. Everything should be an option. Then
again, CodeWright had options for everything and it completely confused me.
The best editor in the world is Brief. Back to DOS!
John Croudy
Alex Bakaev [TeamB] wrote:
> That might have been true of BCB 1.0. Have you tried BCB 3 debugger ?
No, I can't because I can't find a 'demo' version of BCB3. I bought BCB1
after trying the demo version for a few weeks. I don't want to buy BCB3
unless I can test it first. IS there a demo version and I'm missing
something?...
Exception
"Chris Uzdavinis" <chr...@uzdavinis.com> wrote:
* > Ah well... There's no good in discussing preferences. As the Dutch say:
* "Over smaak
* > valt niet te twisten", the Germans say: "Ueber geschmack laBt sich nicht
* reden", or
* > in English: "Beauty is in the eye of the beholder".
*
* Hmmm, that's funny. I've always heard the expression go as, "Beauty is in
* the eye of the person looking at properly aligned columns."
*
* <grin>
Maybe that is in the eye of the beholder of a TListView...
I'm sure you would throw any book away which is completely typeset in Courier.
Grin back, Chris.
Exception
"Chris Uzdavinis" <chr...@uzdavinis.com> wrote:
* > Well, if my programs would have to rely on 'lining up columns of code'
* > for readability, I'd retire! Right now!
*
* I didn't mean for Fortran-like code :) I mean I like it to line up when I
* indent with 2 spaces, the blocks line up. Or maybe this:
*
* int i
* unsigned long x; // this example comment is long and therefore
* // spans 2 lines and I want the comments to be
* aligned.
Bweueeurk. I would typeset these declarations thus:
// Counts the strings.
int i;
// This example comment is long and therefore
// spans 2 lines and I want the comments to be aligned.
unsigned long x;
Why, in the world, would you bother to align the 'i' with the 'x'?
They're totally unrelated in this context, so why line them up?
The character position on which these identifiers begin have nothing to do with their
meaning -- in contrast, proper placed comments and functional empty lines *do* add
meaning to the code, and when thoughtfully placed, can do wonders to readability.
* I'm typing this with proportional fonts, and I want the "i" and "x" to be
* aligned. Can't quite get it. Maybe it's REALLY messed up for some
* people's displays above. The i is approximately above the x on my screen.
* Same with the comments. I want the 2nd row to be directly below the 1st
* row of comments. That will move around on different displays as well.
Lose that stigma of positional obsession. Proportional fonts are more humane.
Therefore, using proportional fonts, a concise writing style is in order.
// If an Abstract Factory is defined...
if (DocFactory) {
DocFactory->Clone(Original); // make a copy
BuildHierarchy(DocFactory); // build the document tree
// Now output it using the chosen format
(Strategy *)OutputFormat->Process(DocFactory->GetResult());
}
The above dummy code fragment is perfectly readable with a proportional font.
* > Oldfashioned monospaced fonts are something of the past, at least for me.
*
* They are cutting-edge technology for me. :)
I notice your smiley, so I don't take your silly statement seriously.
* > * To each his own, I guess. <shrug>
* > That, I agree with you heartily.
*
* I use Emacs and some co-workers use vi. I learned tolerance a long time
* ago through necessity. <g>
If that works in your organisation, ok. But I prefer easy-to-read source code.
Jonathan Arnold
> >>It's not without reason that books are printed in proportional fonts.
> Yeah, but have you ever seen one in which the code examples were in a
> proportional font? ;)
Just one so far, but it's a doozy - Stroustrup's 3rd edition! Here's what
he says:
"At first glance, this presentation style [using proportional-width font]
will seem "unnatural" to programmers accustomed to seeing code in constant-
width fonts. However, proportional-width fonts are generally regarded as
better than constant-width fonts for presntation of text. Using a proportional
width font also allows me to present code with fewer illogical line breaks.
Furthermore, my experiments show that most people find the new style more
readable after a short while."
I'll have to try it with Emacs, I guess.
+===================================================+
| Jonathan Arnold (mailto:jdar...@world.std.com) |
| Programmer Roger Wagner Publishing |
| http://world.std.com/~jdarnold |
+===================================================+
Gregory A. Havenga
Yes, extending the right-click menu would be great.
BUT! I have one peeve about how the editor works - there are no
CUT, COPY, and PASTE selections on the right click menu!!! I have
to either take my hand OFF the mouse and use the ctrl-C/V/X or
shift/ctrl/del combinations, OR, move my mouse 3/4 of the way
across the screen to use the MENU - arghhh! Even if I had the
buttons on the toolbar like in Delphi, I'd still hate it. I don't wanna
move the MOUSE far either!
Peter Steele <pe...@ats-forest.com> wrote in article
<6em4ad$qi...@forums.borland.com>...
> >The tabs at the top of the
> >window are a great idea. I hated having to go to the Window menu in VC.
>
> I would be much happier with the tabs if they were sorted. As it is,
their
> order is essentially random and when I have a lot of files open and some
> tabs are hidden I don't know if I should scroll to the left or scroll to
the
> right to find the file I want. I usually end up going to the list of
files
> in the project manager and clicking on the file there since it is quicker
> than trying to find it on the tabs. Unfortunately, it now looks like the
> files in the project manager are no longer sorted... :-(
>
> What I'd really like is a quick popup menu or pulldown menu right on the
> editor window with a list of open files in it, and another pulldown menu
Kent Reisdorph (TeamB)
>>Just one so far, but it's a doozy - Stroustrup's 3rd edition!
Oops, I have that book, too. Can't say that I agree with him, though.
Proportional _and_ italics. I dunno. I just don't know.
Kent Reisdorph (TeamB)
>>I *hate* clicking on some white
space at then end of the line, and it positioning there.
I actually like that feature. We put our version comments at column 72. It's
nice to be able to click there without inserting a bunch of blank spaces. To
each his own, I guess.
--
Mark G. Wiseman
I have to ask. What is the purpose of putting comments at column 72? Do
you have some kind of automated tool that looks for them there?
Mark
Kent Reisdorph (TeamB) wrote in message <6eosto$6...@forums.borland.com>...
Exception
"Kent Reisdorph (TeamB)" <ke...@turbopower.com> wrote:
* >>It's not without reason that books are printed in proportional fonts.
*
* Yeah, but have you ever seen one in which the code examples were in a
* proportional font? ;)
Pick up Stoustrup 3 and look at the code samples :-)
Mark Cashman
Actually they are not random - I believe they are in the order you
opened them, with a new tab being put next to the currently active tab.
You can exploit this to put things where you want them.
--
------
Mark Cashman, creator of The Temporal Doorway at
http://www.geocities.com/~mcashman
- Original digital art, writing, and more -
C++ Builder Tips and The C++ Builder Programmer's Webring (Join us!)
http://www.geocities.com/~mcashman/progwork.htm
------
Mark Cashman
Good tabs would take care of the alignment problem.
Personally, though, I use the old Borland TE font, which is nice on the
eyes, even when bold, against a black backdrop.
Kent Reisdorph (TeamB)
>>What is the purpose of putting comments at column 72?
Just a code standard. It's this kind of thing (yeah, I know it's Pascal):
Variable : Type; {!!.52}
Can't have those comments all willy-nilly now, can we?<g> We have a coding
standard for obvious reasons, but also because we always ship source code.
Kent Reisdorph (TeamB)
>>IS there a demo version and I'm missing something?...
There's a demo, but it's not a working trial version like BCB 1 had.
Chris Uzdavinis
... except many editors (including Borlands) change spaces/tabs as they see
fit. Emacs uses
<tab> to indent a line, putting its own spacing in appropriately (I love
this). Others see spaces
and replace with tabs, others use tab to align with the token above it.
It's not as simple as using tabs properly. The editor must agree, and the
programmer must also want this.
Do you prefer proportional fonts for source code also? I find this
fascinating.
Chris
Mark Cashman <mcas...@ix.netcom.com> wrote in article
<351016...@ix.netcom.com>...
Chris Uzdavinis
> Bweueeurk. I would typeset these declarations thus:
>
> // Counts the strings.
> int i;
>
> // This example comment is long and therefore
> // spans 2 lines and I want the comments to be aligned.
> unsigned long x;
There have been religious wars regarding whether the * for a pointer should
be
int* x;
or
int *x;
And you're expecting me to just change my style, which is even more
important than
where the * goes? No sirreee, can't do it.
> Why, in the world, would you bother to align the 'i' with the 'x'?
Perhaps I want to keep the variable names easy to read, associated with
their types. When I browse a header file, I look at the members that are
there, THEN I look for its type. It's a lot
easier to read your member variables when they are all in a column.
> They're totally unrelated in this context, so why line them up?
If they're members of a class, they certainly are not totally unrelated.
> Lose that stigma of positional obsession. Proportional fonts are more
humane.
For authors and poets, yes. For programmers, naa.
> * > Oldfashioned monospaced fonts are something of the past, at least
for me.
> *
> * They are cutting-edge technology for me. :)
>
> I notice your smiley, so I don't take your silly statement seriously.
I said it in a joking manner, because obviously monospaced fonts have been
around for
a long time. But they are still important to me.
Chris
Chris Uzdavinis
> No, I can't because I can't find a 'demo' version of BCB3. I bought BCB1
> after trying the demo version for a few weeks. I don't want to buy BCB3
> something?...
They do have a 90-day money back guarentee. So if you like it you're ok,
and if you don't like it you can get your money back, no $$ loss at all.
Just costs
a little time, but so does downloading a demo and installing, uninstalling
that.
But don't worry about returning it, you'll like it. The debugger is almost
worth the
upgrade by itself. :)
Chris
Chris Uzdavinis
If we're talking editor wish lists, I wish the horizontal scrollbar was
proportional to the document width. If my document is not wider than the
screen, the scrollbar should not even be there. If it's barely wider, the
scrollbar should barely move. But as it is, I can scroll thousands of
characters past the right hand side of my document, and the width of the
scrollbar slider is constantly just about an inch. This width should
dynamically adjust, like the vertical scrollbar does.
Yuck.
Chris
Jody Hagins <gam...@magicnet.net> wrote in article
<01bd520e$b53aeec0$39ca06d0@fpfzqlga>...
>
>
> Exception <Exce...@mailcity.com> wrote in article
> <350e40e2...@forums.borland.com>...
> > John Croudy <john....@semitechturku.cod> wrote:
> >
> > * I just thought I'd mention that I think the BCB editor is far better
> than Visual
> > * C++ editor. [...]
> >
>
>
>
> Since we are talking about the editor, I *hate* clicking on some white
> space at then end of the line, and it positioning there. If I click on
> some of the white space area at the end of a line, I want the cursor to
go
> the the last now-whitespace character.
>
Michael Henderson
Mark Cashman wrote in message <351013...@ix.netcom.com>...
>Actually they are not random - I believe they are in the order you
>opened them, with a new tab being put next to the currently active tab.
>
>You can exploit this to put things where you want them.
>
L.J. Wischik
Exception <Exce...@mailcity.com> wrote:
>Why, in the world, would you bother to align the 'i' with the 'x'?
I reckon that most variables are associated, in asmuch as they are all the
local variables for this particular function.
If we had Pascal syntax
var
thisvar : int;
thatvar : long whatever;
pointer : TClass;
then I'd be happy.
As it is, with C++, there's all the
void unsigned static int long
junk before you actually get the the name of the thing you're interested
it.
--
Lucian Wischik, Queens' College, Cambridge CB3 9ET. ljw...@cam.ac.uk
Anduin Withers
All Borland IDE's can be forced to use tabs correctly.
Just some more worthless input on the proportional font topic: I wish
fixed width fonts were used by the debugger. I hate trying to guess if
that j like thing is an i, j, l, t etc are hard to distinguish.
For reading, proportional fonts are great, but for code... well it is
unimaginable. Paste code into word or some other editor that supports
proportional width fonts and argue that it is more readable.
-anduin
Exception
"Kent Reisdorph (TeamB)" <ke...@turbopower.com> wrote:
* I actually like that feature. We put our version comments at column 72. It's
* nice to be able to click there without inserting a bunch of blank spaces. To
* each his own, I guess.
Of course you already knew that there's a handy feature in the code editor which can
help you with this... In Options/Environment/Display, set Display Right Margin to 72.
If you make some edits on a line which makes the comment move, don't you find it
disturbing to have to fix that every time?
John Croudy
Gregory A. Havenga wrote:
> I have to either take my hand OFF the mouse and use the ctrl-C/V/X or
> shift/ctrl/del combinations, OR, move my mouse 3/4 of the way
I use Ctrl/CVX for everything, even in the explorer. Another one of those 'each
to his own' things. Personally, I find it annoying if I have to take my hands
off the *keyboard* because when I'm editing that's what I use. I hate using the
mouse for editing functions. Especially text drag-and-drop. What a horrible
invention. I'm too old for all this. It took me a long time to get used to
having my text in a *window* rather than full-screen. And I prefer fixed width
fonts for programming :)
Kent Reisdorph (TeamB)
>>If you make some edits on a line which makes the comment move, don't you
find it
disturbing to have to fix that every time?
Sure, but that's the price you pay for shipping source with your components.
Kevin Lang
In the old days of FORTRAN, code lines were limited to the first 72
characters on the line. I wonder if this is legacy rearing it's ugly
head??? I know many programming standards limit the length of a line of
code to 80 total characters, so give the developers free reign on the first
72 and let others take control of the last 8...
Personally, I prefer working in landscape and having 132 columns.
Unfortunately, everyone else I know wants to work in portrait, so I have to
adapt. I long for the day when *I'm* dictator :-)
Kevin
Mark G. Wiseman <mwis...@cosolutions.com> wrote in article
<6ep0mc$1g...@forums.borland.com>...
> I have to ask. What is the purpose of putting comments at column 72? Do
> you have some kind of automated tool that looks for them there?
>
> Mark
>
>
> Kent Reisdorph (TeamB) wrote in message
<6eosto$6...@forums.borland.com>...
> >I actually like that feature. We put our version comments at column 72.
> It's
> >nice to be able to click there without inserting a bunch of blank
spaces.
> To
Exception
"Kevin Lang" <kjl...@spambasket.gcs-server.jsc.nasa.gov> wrote:
* Personally, I prefer working in landscape and having 132 columns.
* Unfortunately, everyone else I know wants to work in portrait, so I have to
* adapt. I long for the day when *I'm* dictator :-)
Why should every programmer in a team apply the same coding standard, bracketing,
variable naming rules, layout, etc... etc... Makes me think of a Communist Party
which imposes its standards upon the masses, crushing any creative spike there may
exist. I wouldn't like to have a job in such an environment. I dislike stupid rules.
Developers in a team could use the time they spend in endless debates over where
braces should be placed, or how variables should be spelled, better for getting their
interfaces right.
Kevin Lang
> "Kevin Lang" <kjl...@spambasket.gcs-server.jsc.nasa.gov> wrote:
>
> * Personally, I prefer working in landscape and having 132 columns.
> * Unfortunately, everyone else I know wants to work in portrait, so I
have to
> * adapt. I long for the day when *I'm* dictator :-)
>
> Why should every programmer in a team apply the same coding standard,
bracketing,
> variable naming rules, layout, etc... etc... Makes me think of a
Communist Party
> which imposes its standards upon the masses, crushing any creative spike
there may
> exist. I wouldn't like to have a job in such an environment. I dislike
stupid rules.
>
> Developers in a team could use the time they spend in endless debates
over where
> braces should be placed, or how variables should be spelled, better for
getting their
> interfaces right.
I think a prime reason for it is that developers don't always finish a
project they start on. If everyone adheres to a common style guide, it
should be easier for another person to pick up. I don't think that coding
standards stifle creativity, just make code reviews and other QA and
documentation processes go a lot smoother. When you code within a larger
organization, it's desirable for the code to look like a single entity
developed it.
I don't think that, organizationally, it matters whether proportional or
fixed fonts are used, or whether all local variables are declared at the
top of a function block, or whatever. It does matter if 20 developers are
each presenting their code in 20 different ways with 20 different naming
standards, and 20 different strategies for variable declarations. What you
look at on screen, and what you present to others for review can be two
different things, but everything in life requires some responsibility to
temper personal tastes with group commonalities. Otherwise, more people
would be wearing striped shirts with pocket protectors coordinating nicely
with plaid pants. :-) That's why today's nerd wears cut-off Levi's and
t-shirts.
Kevin
Chris Uzdavinis
> Why should every programmer in a team apply the same coding standard,
bracketing,
> variable naming rules, layout, etc... etc... Makes me think of a
Communist Party
> which imposes its standards upon the masses, crushing any creative spike
there may
> exist. I wouldn't like to have a job in such an environment. I dislike
stupid rules.
While I agree with you in spirit, it seems that the hardest part about
moving to a new project is finding multiple convlictin naming conventions,
layout schemas, indentation, bracket placement... it is hard to get
continuity anywhere. If it all looks the same, you get into the right
mindset. You don't have to remember that person X always puts global
variables in namespace global, and person Y puts them as static members of
a class. If functions always start with uppercase letters and variables
with lower case letters, you know something about an identifier without
having to look it up.
Documentation generators seem to work better when the code style is the
same too.
> Developers in a team could use the time they spend in endless debates
over where
> braces should be placed, or how variables should be spelled, better for
getting their
> interfaces right.
I have a style that I hate to deviate from and I cringe when I see others
veer off the "true path" (mine). The hardest thing is to have to conform
to someone else's style by joining their team. I'd like to start my own
business, and then dictate my style onto others for once!
Chris
Jerry Bloomfield
On Tue, 17 Mar 1998 21:56:39 GMT, Exce...@mailcity.com (Exception)
wrote:
>"Chris Uzdavinis" <chr...@uzdavinis.com> wrote:
>
>* Proportional fonts? For programs? Whoa, I thought I had heard it all!
>* Personally, I can't stand proportional fonts with code because nothing
>* lines up correctly and everything is difficult to read.
>
>Well, if my programs would have to rely on 'lining up columns of code' for
>readability, I'd retire! Right now!
Actually, you would probably retire, and then take a consulting job
as a Y2K analyst making about twice what you make now to read that old
COBOL code and tell them it will or won't work right in a couple years
<g>
Jerry Bloomfield
On Wed, 18 Mar 1998 10:04:10 GMT, Exce...@mailcity.com (Exception)
wrote:
>Therefore, using proportional fonts, a concise writing style is in order.
>
> // If an Abstract Factory is defined...
> if (DocFactory) {
> DocFactory->Clone(Original); // make a copy
> BuildHierarchy(DocFactory); // build the document tree
> // Now output it using the chosen format
> (Strategy *)OutputFormat->Process(DocFactory->GetResult());
> }
>
>The above dummy code fragment is perfectly readable with a proportional font.
Well, other than the style issues of how you should align your braces
;-) I believe the commonly accepted response to that is that your
"style" should match the rest of the code in which you are working to
maintain readability. Since you have the luxury of working on your
own code, then the proportions (and braces <g>) meet your style.
Unfortunately most of the rest of us have old code we must still
maintain as well, and to maintain the readability, we can't just
change things like that in mid-program...
>I notice your smiley, so I don't take your silly statement seriously.
We should never take any statements made here too seriously, smiley
or not... If we do, we will no doubtfully get pissed about something
sooner or later...
Michael Henderson
Exception wrote in message <35117ca0...@forums.borland.com>...
>
>Why should every programmer in a team apply the same coding standard,
bracketing,
>variable naming rules, layout, etc... etc... Makes me think of a Communist
Party
>which imposes its standards upon the masses, crushing any creative spike
there may
>exist. I wouldn't like to have a job in such an environment. I dislike
stupid rules.
>
actually increase if everyone is using a common system. it takes less time
to pick up new code, makes the code reviews and walkthroughs easier. from a
management perspective, you want code that is maintainable and can be
evaluated from "outside" of the team.
as a consultant, i have to adhere to the client's guidelines. if i'm unlucky
enough to work in a place where the management cut they're teeth on FORTRAN
or COBOL, i'm going to spend a lot of time placing comments in column X. i
used to chafe against this because i learned most of what i know working in
a small shop in alaska. we could tell who'd written which piece of code just
by the way it looked. that worked fairly well (even if we were continually
moving each other's curly braces :)
that attitude changed after i went to a few places that used a more
egalitarian philosophy (everyone used whatever they wanted). that ran the
gamut from no indentation for anything to programs that could compete in
calligraphy contests (god forbid that the non-indenter do maintenance on one
of those). it also let people get away with using crappy names (i can't
forget the guy who used (and reused) the globals "t1, t2, t3, t4, t5, t7,
t8, t9" we could never figure out why he didn't like the number six). it was
tough working with those environments since i couldn't get clues from the
program structure and the useless variable names. that meant more learning
time and a strong dislike for the term "self documenting code."
after that, if i worked at a place w/o standards, i'd propose some. usually
wasn't too hard to convince the management that the time spent was worth it.
Exception
Jers...@wwa.com (Jerry Bloomfield) wrote:
* Exception wrote:
*
* >Chris Uzdavinis wrote:
* >
* >* Proportional fonts? For programs? Whoa, I thought I had heard it all!
* >* Personally, I can't stand proportional fonts with code because nothing
* >* lines up correctly and everything is difficult to read.
* >
* >Well, if my programs would have to rely on 'lining up columns of code' for
* >readability, I'd retire! Right now!
*
* Actually, you would probably retire, and then take a consulting job
* as a Y2K analyst making about twice what you make now to read that old
* COBOL code and tell them it will or won't work right in a couple years
* <g>
Good idea. Finding dates in programs can't be that difficult, even if its COBOL
(haven't seen that since university). But how am I going to spend that $20.000 a
month.
And then, in a year of two, when all that software collapses and they're out to get
me, I'll be sitting on a quiet Caribbean Island, with a new identity. Juice!
Michael Henderson
Exception wrote in message <35129e53...@forums.borland.com>...
>
>Good idea. Finding dates in programs can't be that difficult, even if its
COBOL
>
had a date <g>
Edward Diener
Except that if you have to deal with Cobol in any way you might very well have a heart
attack before you get to bask on that Caribbean island.
Exception wrote:
> Jers...@wwa.com (Jerry Bloomfield) wrote:
>
> * Exception wrote:
> *
> * >Chris Uzdavinis wrote:
> * >
> * >* Proportional fonts? For programs? Whoa, I thought I had heard it all!
> * >* Personally, I can't stand proportional fonts with code because nothing
> * >* lines up correctly and everything is difficult to read.
> * >
> * >Well, if my programs would have to rely on 'lining up columns of code' for
> * >readability, I'd retire! Right now!
> *
> * Actually, you would probably retire, and then take a consulting job
> * as a Y2K analyst making about twice what you make now to read that old
> * COBOL code and tell them it will or won't work right in a couple years
> * <g>
>
> Good idea. Finding dates in programs can't be that difficult, even if its COBOL
Jody Hagins
Kent Reisdorph (TeamB) <ke...@turbopower.com> wrote in article
<6erh8m$2m...@forums.borland.com>...
> >>If you make some edits on a line which makes the comment move, don't
you
> find it
> disturbing to have to fix that every time?
>
> Sure, but that's the price you pay for shipping source with your
components.
I've never found that putting comments at the end of lines was very
readable. I prefer to put them on lines by themselves, with the same
indentation as the code block
while (someCondition)
{
// Blah blah blah
doSomething
if (someCondition)
{
// Some more blathering
some code
}
}
Jody Hagins
Exception <Exce...@mailcity.com> wrote in article
<35117ca0...@forums.borland.com>...
> "Kevin Lang" <kjl...@spambasket.gcs-server.jsc.nasa.gov> wrote:
>
> * Personally, I prefer working in landscape and having 132 columns.
> * Unfortunately, everyone else I know wants to work in portrait, so I
have to
> * adapt. I long for the day when *I'm* dictator :-)
>
> Why should every programmer in a team apply the same coding standard,
bracketing,
> variable naming rules, layout, etc... etc... Makes me think of a
Communist Party
> which imposes its standards upon the masses, crushing any creative spike
there may
> exist. I wouldn't like to have a job in such an environment. I dislike
stupid rules.
>
> Developers in a team could use the time they spend in endless debates
over where
> braces should be placed, or how variables should be spelled, better for
getting their
> interfaces right.
Well, there are a few points here...
If you are on a team, use the indent program. Whenever anything gets
thrown into the revision control system, indent it with all defaults,
before throwing it in (this keeps diff from going nuts). The when you
check it out, run indent on it again with your parameters, and wala, your
code now adheres to your own formatting. It's quite simple for revision
control systems that have checkin/checkout hooks, since you just modify the
hook, and you never even know that your code is stored in the RCS
differently. When each developer looks at it, the code will be in their
format of choice.
For those who don't know what indent is, it's a C/C++ code formatter. I'm
sure there are PC versions running around somewhere.
For variables, I don't think naming matters much since they should be
shielded within each class, and you shouldn't have behemoth functions
anyway. However, I *do* think it matters for class methods. For example,
how are getters/setters implemented (references, get/set, overloaded name).
Also, naming convention for classes can get confusing if all classes have
different ways of getting at data. That's one place it should be standard.
Anyway, for those interested, here's what I do (though I used to overload
getters/setters... the Smalltalker in me I then switched to a hybrid, and
now I use get/set... the java in me)
class Foo
{
public:
enum SomeEnum
{
se_EnumsStartWithIdTagFollowedBySomethingLikeThis
};
int instanceMethodsStartWithLowerCase();
static int StaticMethodsStartWithUpperCase();
Something const & getSomething() const;
void setSomething(Something const &);
private:
int instanceMembersStartWithLowerCaseAndEndInUnderscore_;
static int StaticMembersStartWithUpperCaseAndEndInUnderscore_;
};
I like lining up braces, and *always* putting in braces, indenting 2 spaces
for (int i = 0; i < someValue; ++i)
{
doSomething();
}
Also, I separate class methods on different lines like:
int Foo::
instanceMethodsStartWithLowerCase()
{
}
This is because under vi, I can go to functions easily using search for
symbol at beginning of line. It's easier on the eyes too (well *my* eyes
anyway).
Can;t really think of anything else (though I could probably see it in
code...
Kent Reisdorph (TeamB)
>>I've never found that putting comments at the end of lines was very
readable.
We only do that for the version number comments.
Mark Cashman
You can if you are just trying to get the header and cpp next to each
other. If you are in the header or cpp, pick open header / source from
the popup, and open or not already, it opens next to the current window.
At least its a partial solution.
--
------
Mark Cashman, creator of The Temporal Doorway at
http://www.geocities.com/~mcashman
- Original digital art, writing, and more -
C++ Builder Tips and The C++ Builder Programmer's Webring (Join us!)
http://www.geocities.com/~mcashman/progwork.htm
------
Francis Smit
My problem is thus, I use DataModules for a bunch of components {TDatabase,
TTable's, TQuery's, TImageList's}, I refer to these from a form, every time I
start builder, I find them all disconnected, and I have to redo the whole thing.
I remember Delphi 1.0 did something like this but just on a form, the tables/queries
would come up, and they'd be inactive even tho' you'd saved them as active; I
think the fix was to change the creation order.
I've tried that here no luck, this only happens at design time??, what's the fix??, is
this
a bug??, why doesn't builder take care of these dependencies for me???.
Hope borland are listening, may be I should mail them as well!!!