On Thu, Jan 01, 2015 at 06:53:00PM -0600, Larry Garfield wrote:

> The "end users" of php-src are "people who write PHP code".  Those
> are the end users that we should be concerned with.  "People who
> visit web sites" are *their* end users.  Those people don't care in
> the slightest what happens on this list; they care that the people
> writing PHP code can do their job in a minimum amount of time and
> with a minimum amount of bugs.
> 
> So asking developers of the major PHP frameworks and applications
> what would help them do their job in a minimum amount of time with a
> minimum amount of bugs is absolutely a worthwhile endeavor to figure
> out what would be "best".

I beleive that you are making a mistake by assuming that all our 'users' are the
same and have the same goals. There are:

a) those producing a web site, want to do so quickly, are not really concerned
about occasional type errors (maybe they should, but that is another story) and
for who current weak typing (inc type juggling) is just what they want.  The
application prob had little real design but runs to their satisfaction, prob has
few comment or other documentation.

b) those who are producing a package/framework which takes sloppy input (from
forms, etc), validate it and thereafter variables should contain known types.
Also here are those who do like to remove all errors from even small 
applications.
They might not care if a value is int 42 or string "42", but want to be told (ie
error) if it is "42 fish" or "fourty two". They can then fix their validation.
They want to do the validation in one/few places, not have to put checks all 
over.
These users are willing to put in the effort to get it right, they have designed
it and will document it, etc.

There are many in between (a) & (b) in various combinations.

So, the needs/wants of (a) and (b) is very different.

IMHO part of the reason for the high contention is that people comment from the
perspective of either (a) or (b) and thus will not agree with someone from the
other camp.

Part of the solution is making type checking optional - as is the proposal.
But do try to remember that we do have different types of user.


Me ? I am in the (b) camp but do occasionally stray to (a).

There is a 3rd camp: (c)
These are the language implementors, in particular 3rd parties eg HipHop. These
do care if a variable is int 42 or string "42" because there are optimisations
to be made if they know that a value is really an int and not a string cleanly
convertable to an int.

-- 
Alain Williams
Linux/GNU Consultant - Mail systems, Web sites, Networking, Programmer, IT 
Lecturer.
+44 (0) 787 668 0256  http://www.phcomp.co.uk/
Parliament Hill Computers Ltd. Registration Information: 
http://www.phcomp.co.uk/contact.php
#include <std_disclaimer.h>

-- 
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php

Reply via email to