Thread overview
deprecating the body keyword
Dec 08, 2009
hehe45
Dec 08, 2009
Bill Baxter
Dec 08, 2009
retard
Dec 08, 2009
Lutger
Dec 09, 2009
retard
Dec 09, 2009
Jesse Phillips
December 08, 2009
I haven't been following the newgroup closely, so I don't know if this has already been discussed, but I wanted to make a few suggestions before D2 is finalized.
I think the "body" keyword is completly useless unless it is required because of parsing issues.
It would also be great if the order of the "in" and "out" contracts would not be fixed by the compiler, so they could be written in any order around the function body, so one could put the "out" contract after the function for example.
These improvements could be added to D1 too.
December 08, 2009
On Tue, Dec 8, 2009 at 2:11 PM, hehe45 <a3161739@uggsrock.com> wrote:
> I haven't been following the newgroup closely, so I don't know if this has already been discussed, but I wanted to make a few suggestions before D2 is finalized.
> I think the "body" keyword is completly useless unless it is required because of parsing issues.

Search for "I hate my body".
That was the title of the thread started a while back.


--bb
December 08, 2009
Tue, 08 Dec 2009 14:59:10 -0800, Bill Baxter wrote:

> On Tue, Dec 8, 2009 at 2:11 PM, hehe45 <a3161739@uggsrock.com> wrote:
>> I haven't been following the newgroup closely, so I don't know if this has already been discussed, but I wanted to make a few suggestions before D2 is finalized. I think the "body" keyword is completly useless unless it is required because of parsing issues.
> 
> Search for "I hate my body".
> That was the title of the thread started a while back.

Yes, yet another thread started but basically nothing happened. How frustating it is that Walter never expresses his intentions before a new release? This kind of mentality is from early 20th century. Nowadays all agile projects clearly define the goals for the next sprint.
December 08, 2009
retard wrote:

> Tue, 08 Dec 2009 14:59:10 -0800, Bill Baxter wrote:
> 
>> On Tue, Dec 8, 2009 at 2:11 PM, hehe45 <a3161739@uggsrock.com> wrote:
>>> I haven't been following the newgroup closely, so I don't know if this has already been discussed, but I wanted to make a few suggestions before D2 is finalized. I think the "body" keyword is completly useless unless it is required because of parsing issues.
>> 
>> Search for "I hate my body".
>> That was the title of the thread started a while back.
> 
> Yes, yet another thread started but basically nothing happened. How frustating it is that Walter never expresses his intentions before a new release? This kind of mentality is from early 20th century. Nowadays all agile projects clearly define the goals for the next sprint.

That's usually more of a business thing though, when you write the features other people prioritize.

Besides, image Walter responding to all requests, these threads generally explode pretty fast. How much time do you think he will have left to actually work on the compiler? I wonder how Andrei manages his level of participation in the newsgroups.


December 09, 2009
Wed, 09 Dec 2009 00:44:24 +0100, Lutger wrote:

> retard wrote:
> 
>> Tue, 08 Dec 2009 14:59:10 -0800, Bill Baxter wrote:
>> 
>>> On Tue, Dec 8, 2009 at 2:11 PM, hehe45 <a3161739@uggsrock.com> wrote:
>>>> I haven't been following the newgroup closely, so I don't know if this has already been discussed, but I wanted to make a few suggestions before D2 is finalized. I think the "body" keyword is completly useless unless it is required because of parsing issues.
>>> 
>>> Search for "I hate my body".
>>> That was the title of the thread started a while back.
>> 
>> Yes, yet another thread started but basically nothing happened. How frustating it is that Walter never expresses his intentions before a new release? This kind of mentality is from early 20th century. Nowadays all agile projects clearly define the goals for the next sprint.
> 
> That's usually more of a business thing though, when you write the features other people prioritize.
> 
> Besides, image Walter responding to all requests, these threads generally explode pretty fast. How much time do you think he will have left to actually work on the compiler? I wonder how Andrei manages his level of participation in the newsgroups.

I don't expect him to respond to all requests, most of the new threads are silly bikeshedding. But there are often megathreads with 500+ posts. If some feature seems so important to the majority and especially to veteran users, I expect some kind of opinion also from the language's author. I expect the Andrei's book to have a major effect on the release date of D2. A clear roadmap is definitely something that would help now.

You somehow seem to assume that his only work is to write the compiler. As an author of the language he indeed also has to make all kinds of design decisions. In the end it is Walter who's in charge of everything. Andrei's participation surely helps, he has lots of experience and brain capacity, often also great, perhaps even radical, new ideas. Walter never expresses his opinions nor intentions. I cannot know by inspecting dmd what d actually should do since those features can also be bugs.
December 09, 2009
retard wrote:

> I don't expect him to respond to all requests, most of the new threads are silly bikeshedding. But there are often megathreads with 500+ posts. If some feature seems so important to the majority and especially to veteran users, I expect some kind of opinion also from the language's author. I expect the Andrei's book to have a major effect on the release date of D2. A clear roadmap is definitely something that would help now.

Those threads of 500+ posts are usually because the community doesn't
have consensus on what to do. Though if Walter/Andrei are the
only ones against the idea then the thread could be that large. Though
many of the ideas in those threads get snuck into a later release.

I think it is good that Walter pushes back on the community as hard as he does. It of course would be nice if final decision wasn't packed as a surprise for the next release (now commit). Who knows maybe next release will have non-nullable types as the default.

Personally I don't have an issue with the body keyword, nor that order is mandated. Having a random order to in, out seems pointless.