1. Private email is private discussion and should stay that way.

2. Public forum messages (and especially here) may be copied, reused, etc.

3. The interfaces topic is of interest to many in this forum.  Whether or not there is private email negotiating the needs and ways to solve them, I think there does need to be a policy level announcement by Walter on this topic here. Too many questions have been opened, opinions stated, problems described, to just leave things the way they have been.

4. It has been a hot day in spite of the chill and rain here in PA.

5. Limbo from AT&T sounds even more like D than Java or C++ does, but it's interpreted...

"larry cowan" <larry_member@pathlink.com> wrote in message news:c55fbd$107b$1@digitaldaemon.com...
> 3. The interfaces topic is of interest to many in this forum.  Whether or
not
> there is private email negotiating the needs and ways to solve them, I
think
> there does need to be a policy level announcement by Walter on this topic
here.
> Too many questions have been opened, opinions stated, problems described,
to
> just leave things the way they have been.

Actually, I've been working on a solution, but I want to make sure it works correctly first.

Thanks.  I didn't mean instantly, but we need some future plans with more than a few words.  It is a a large concern to many of us, and to you, I'm sure.  The commercial useability of D is very important to us also.

In article <c55klr$19bd$1@digitaldaemon.com>, Walter says...
>
>
>"larry cowan" <larry_member@pathlink.com> wrote in message news:c55fbd$107b$1@digitaldaemon.com...
>> 3. The interfaces topic is of interest to many in this forum.  Whether or
>not
>> there is private email negotiating the needs and ways to solve them, I
>think
>> there does need to be a policy level announcement by Walter on this topic
>here.
>> Too many questions have been opened, opinions stated, problems described,
>to
>> just leave things the way they have been.
>
>Actually, I've been working on a solution, but I want to make sure it works correctly first.
>
>

I hate to say I told you so, but:
	I told you so.

Go get 'em, Walter.  Looking for an employee, by any chance? :)

Walter wrote:
> Actually, I've been working on a solution, but I want to make sure it works
> correctly first.

"Russ Lewis" <spamhole-2001-07-16@deming-os.org> wrote in message news:c56fc5$2j72$2@digitaldaemon.com...
> Go get 'em, Walter.  Looking for an employee, by any chance? :)

Yes, one who will work for free <g>.

larry cowan said:
> Thanks.  I didn't mean instantly, but we need some future plans with more than a
> few words.  It is a a large concern to many of us, and to you, I'm sure.  The
> commercial useability of D is very important to us also.
> 
> In article <c55klr$19bd$1@digitaldaemon.com>, Walter says...
> 
>>
>>"larry cowan" <larry_member@pathlink.com> wrote in message
>>news:c55fbd$107b$1@digitaldaemon.com...
>>
>>>3. The interfaces topic is of interest to many in this forum.  Whether or
>>
>>not
>>
>>>there is private email negotiating the needs and ways to solve them, I
>>
>>think
>>
>>>there does need to be a policy level announcement by Walter on this topic
>>
>>here.
>>
>>>Too many questions have been opened, opinions stated, problems described,
>>
>>to
>>
>>>just leave things the way they have been.
>>
>>Actually, I've been working on a solution, but I want to make sure it works
>>correctly first.
>>
>>
> 
> 
> 
I totally feel your worrying of the future of D interface. I am now starting looking at D and see that just a Interface could case so many problems, I am more worrying using D or not.

Just take a look at something like java. We all know that java people use interface very much, and we know that they also really "implements" the interface. I couldn't find that happening, even in future, in D. Just only the syntax is confusing, a ":" stands for "implements", but it also stand for "inherits", so which is the real thing? Without a rough look of the original code of the interface implemented, we could not see if it is a interface or not.

Also as others here said, is D interface really a interface? I would not say it is. From other examples, I can only state a use of interface in D -- mapping a COM object. I use "map" because it is really a mapping, and we know that interface originally is designed for the "maker", not the user. What I want to point out is that "interface" is not a wrapper for user of the class/object, but for class maker to give a better organized front-end to users. I would suggest, renaming the keyword interface to wrapper, which is a reflection of current use.

In the other hand, the real interface usage stated above could be incorporated as "properties". It is actually from C#. From other forums and mailing-list or etc., I get the idea that properties is actually a improved version of interfaces. In my point of view properties are readable and/or writable, so does interfaces. Of course they are not totally the same, but if it is possible properties would give better results than interfaces.

After all, D *is* D. It is not C, VC, Java, C# etc. There is no need for D to copy others, but there is a need to make things as clear as possible. I hope somebody here would give brighter ideas than annoying noises.

"School" <tec_mailhk@yahoo.com.hk> wrote in message news:c56p63$2un$1@digitaldaemon.com...
> I totally feel your worrying of the future of D interface. I am now starting looking at D and see that just a Interface could case so many problems, I am more worrying using D or not.
>
And the good news is ... Walter has resolved all those issues!

- Kris

Kris ??:
> "School" <tec_mailhk@yahoo.com.hk> wrote in message
> news:c56p63$2un$1@digitaldaemon.com...
> 
>>I totally feel your worrying of the future of D interface. I am now
>>starting looking at D and see that just a Interface could case so many
>>problems, I am more worrying using D or not.
>>
> 
> And the good news is ... Walter has resolved all those issues!
> 
> - Kris
> 
> 
What a good news!

Just a comment... you're thinking exclusively about dispinterfaces (or properties/methods called through an IDispatch interface)... full COM Interfaces are used by clientes (i.e. IDispatch which is used to set/get properties and call methods of dispinterfaces)

> larry cowan said:
> > Thanks.  I didn't mean instantly, but we need some future plans with
more than a
> > few words.  It is a a large concern to many of us, and to you, I'm sure.
The
> > commercial useability of D is very important to us also.
> >
> > In article <c55klr$19bd$1@digitaldaemon.com>, Walter says...
> >
> >>
> >>"larry cowan" <larry_member@pathlink.com> wrote in message news:c55fbd$107b$1@digitaldaemon.com...
> >>
> >>>3. The interfaces topic is of interest to many in this forum.  Whether
or
> >>
> >>not
> >>
> >>>there is private email negotiating the needs and ways to solve them, I
> >>
> >>think
> >>
> >>>there does need to be a policy level announcement by Walter on this
topic
> >>
> >>here.
> >>
> >>>Too many questions have been opened, opinions stated, problems
described,
> >>
> >>to
> >>
> >>>just leave things the way they have been.
> >>
> >>Actually, I've been working on a solution, but I want to make sure it
works
> >>correctly first.
> >>
> >>
> >
> >
> >
> I totally feel your worrying of the future of D interface. I am now starting looking at D and see that just a Interface could case so many problems, I am more worrying using D or not.
>
> Just take a look at something like java. We all know that java people use interface very much, and we know that they also really "implements" the interface. I couldn't find that happening, even in future, in D. Just only the syntax is confusing, a ":" stands for "implements", but it also stand for "inherits", so which is the real thing? Without a rough look of the original code of the interface implemented, we could not see if it is a interface or not.
>
> Also as others here said, is D interface really a interface? I would not say it is. From other examples, I can only state a use of interface in D -- mapping a COM object. I use "map" because it is really a mapping, and we know that interface originally is designed for the "maker", not the user. What I want to point out is that "interface" is not a wrapper for user of the class/object, but for class maker to give a better organized front-end to users. I would suggest, renaming the keyword interface to wrapper, which is a reflection of current use.
>
> In the other hand, the real interface usage stated above could be incorporated as "properties". It is actually from C#. From other forums and mailing-list or etc., I get the idea that properties is actually a improved version of interfaces. In my point of view properties are readable and/or writable, so does interfaces. Of course they are not totally the same, but if it is possible properties would give better results than interfaces.
>
> After all, D *is* D. It is not C, VC, Java, C# etc. There is no need for D to copy others, but there is a need to make things as clear as possible. I hope somebody here would give brighter ideas than annoying noises.

Actually, a com interface is a MS C++ compatible virtual pointer table. It is defilned that way to be completely simple to implement in MS C++ (et al). It is a binary standard for defining functions (methods) and not properties. COM requires that all interfaces derive from IUnknown; however, this is just a way of providing object life time control by reference counting.

The IDispatch inteface is another layer on top of this, implementing a bunch of functions that allow 'late bound' applications to pass all functions calls (and property set/reads) through a single function as varaints - very ugly.

The ultimate is to build on top of an IDispatch interface and get both the 'late bound' capability as well as the direct call capability via the vtbl's pointers to functions. The standard is to then define properties as both get and set/put methods on the interface - so each property requires two functions one to read and one to write.

The meta data about the inteface (type library) is used by the standard COM IDispatch implementation to pass the variant based calls to the actual vtbl ones. An interface defiend this way is called 'dual'.

Does this make sence to anyone - does anyone care?



"School" <tec_mailhk@yahoo.com.hk> wrote in message news:c56p63$2un$1@digitaldaemon.com...
> larry cowan said:
> > Thanks.  I didn't mean instantly, but we need some future plans with
more than a
> > few words.  It is a a large concern to many of us, and to you, I'm sure.
The
> > commercial useability of D is very important to us also.
> >
> > In article <c55klr$19bd$1@digitaldaemon.com>, Walter says...
> >
> >>
> >>"larry cowan" <larry_member@pathlink.com> wrote in message news:c55fbd$107b$1@digitaldaemon.com...
> >>
> >>>3. The interfaces topic is of interest to many in this forum.  Whether
or
> >>
> >>not
> >>
> >>>there is private email negotiating the needs and ways to solve them, I
> >>
> >>think
> >>
> >>>there does need to be a policy level announcement by Walter on this
topic
> >>
> >>here.
> >>
> >>>Too many questions have been opened, opinions stated, problems
described,
> >>
> >>to
> >>
> >>>just leave things the way they have been.
> >>
> >>Actually, I've been working on a solution, but I want to make sure it
works
> >>correctly first.
> >>
> >>
> >
> >
> >
> I totally feel your worrying of the future of D interface. I am now starting looking at D and see that just a Interface could case so many problems, I am more worrying using D or not.
>
> Just take a look at something like java. We all know that java people use interface very much, and we know that they also really "implements" the interface. I couldn't find that happening, even in future, in D. Just only the syntax is confusing, a ":" stands for "implements", but it also stand for "inherits", so which is the real thing? Without a rough look of the original code of the interface implemented, we could not see if it is a interface or not.
>
> Also as others here said, is D interface really a interface? I would not say it is. From other examples, I can only state a use of interface in D -- mapping a COM object. I use "map" because it is really a mapping, and we know that interface originally is designed for the "maker", not the user. What I want to point out is that "interface" is not a wrapper for user of the class/object, but for class maker to give a better organized front-end to users. I would suggest, renaming the keyword interface to wrapper, which is a reflection of current use.
>
> In the other hand, the real interface usage stated above could be incorporated as "properties". It is actually from C#. From other forums and mailing-list or etc., I get the idea that properties is actually a improved version of interfaces. In my point of view properties are readable and/or writable, so does interfaces. Of course they are not totally the same, but if it is possible properties would give better results than interfaces.
>
> After all, D *is* D. It is not C, VC, Java, C# etc. There is no need for D to copy others, but there is a need to make things as clear as possible. I hope somebody here would give brighter ideas than annoying noises.