Re: Preview: portable dumper

[Paul Eggert]

On 11/29/2016 11:35 AM, Eli Zaretskii wrote:

> I didn't compare the proposed portable dumper with what we have now,
That's our fundamental disagreement, then. I am willing to settle for 
real improvements to Emacs even if they are not ideal. And I think that 
this sort of willingness is more likely to attract contributions from 
developers.

> leading a project to be viable in the
> future needs to look at the reality, not at our wishful thinking
Daniel's proposed change consists of code that works and that runs fast. 
Isn't that reality? And in contrast, isn't the lread-based approach 
wishful thinking?

> When did you last try to hire a C or C++ programmer?

I don't *hire* them, I *create* them. I currently have 450 undergraduate 
students, and C++ and C are the only programming languages that they 
have all written and debugged code in. (This is not just my decision; it 
comes from our faculty's curriculum.) According to Philip Guo's 2014 
survey, the most popular languages used in teaching computer science in 
top U.S. universities are Python (25-30%), Java (20-25%), C/C++ 
(10-15%), and MATLAB (5-10%). So from what I can see, the C pipeline is 
by no means empty.

Anyway, although C has probably passed its peak and will surely die out 
eventually, it's unlikely for this to be anytime soon. The prospect of 
its demise shouldn't materially affect our choice of Emacs dumping 
technology in the meantime.