[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

if (c++ && x++)

Responses to:

Hugh & Roger -- I agree that there are lots of cool Developers 
who will see the transition to X++ as a minor detour in a Journey 
upon which they've already embarked. We will seek these folks 
out and give them our X++ tools and they'll provide us with the 
potential for some great applications.

But if we want Xanadu to become anything more than an interesting 
niche technology being used and exercised by the programming 
elite, I believe that we have to present the technology to the 
widest possible audience as early as possible.

Note that we're in substantial agreement -- my issue has to do 
with timing and emphasis. The Developer community will form its 
initial impressions of our technology based on what we first 
release, and any attempt to alter those first impressions will 
meet with the same sort of resistance that's greeted MacApp over 
the **YEARS** that Apple's been trying to promote it (sluggish, 
NIH, requires learning Object Pascal, etc.). I believe it's critical 
to our success that Developers think of the Xanadu development 
environment *from the beginning* as one directly and easily accessible 
from C, with wonderful and powerful extensions available for 
C++ programmers (including Compositors, FM, etc.). This is my 
main point, and I don't mind undertaking the work to support 
both environments (this part of message will self-destruct after 
you've read it and I'll deny ever having made that statement...:-).

I think it would be mistake to seal off the C libraries or to 
cripple them in any obvious way. Over time we can achieve the 
same effect by adding neat new stuff to the X++ libraries and 
only selective portions available from C (which portions is another 
problem). My general feeling is that we should let natural forces 
dictate which directions we pursue, rather than force our particular 
flavor of the future on the developer community.

Dean -- "Fearsome as that is, starting from scratch [to document 
X++] is probably much easier than dealing with C flavored C++". 
Yes, but easier for whom? I don't think we should make our decisions 
about the level of support we provide for either environment 
on the basis of how much work it is for us (although it's of 
course important). It's equally important to consider the perception 
Developers will have of how much work is required of THEM. And 
this is the broader issue I'm concerned with -- how we position 
this product (which is, after all, essentially a development 
environment). In my opinion, we *don't* want Xanadu to be viewed 
as primarily a C++ environment. It must be positioned as an environment 
equally accessible from C or C++. Developers who choose to use 
C++ will have X++ and other significant advantages but vanilla 
C programmers should not feel that they are second class citizens 
in the Xanadu community.

Again, it's mostly a question of shifting emphasis and timing. 
Developers who call with their C questions shouldn't be told 
"you should be using C++". And if I had to choose which materials 
were available first (and I may have to), I'd choose the C materials 
and let the C++/X++ materials follow. Both must be done and supported, 
but in my view initial acceptance by the C community is much 
more important to our eventual successes.

Roger -- As I said above, we're in substantial -- even violent 
-- agreement. But I do disagree with your assertion that C programmers 
choose C primarily because of their dependence on existing libraries. 
There's also a significant number of C programmers who would 
prefer to develop in C simply *because they already know C*. 
That's a substantial advantage for busy people who have tools 
that they know already serve them well.

-- bobp