The issue I have with PHP is that the people in charge have reasons for not
implementing performance enhancements in the base code. They charge a fair
amount for add-ons that increases performance drastically. I could
actually argue that extensibility and performance on the back end aren't
what they should be for this reason.
Not that I want to make enemies here, but I think this is a realistic
criticism. Not to mention that the Qt license that is used prevents anyone
from making extensions and selling them w/o an additional license from the
Zend people. So they are able to make money off of the hard work of all of
the module contributors, which I think really blows.
I really enjoy using PHP. I think the authors have done a commendible job.
I just wish that it was more open.
Medvitz
Ken Egervari wrote:
> Although your arguments make sense for speed, this is tradeoff that many
> programmers are willing to take. As for taking "tons of time to load",
> although I have noticed a large slowdown, it's not critical and nothing a
> better server can't solve if it does become critical.
>
> I'd rather develop a website in half the time and spend more money on a
> server than do it slower and harder. No one wants to work hard.
>
> As for the data layer, I think simple calls like that don't constitute a
> data layer at all. You still might have database code all over the
> website, and many of the related things like
> adding/updating/deleting/searching/whatever on a single entity can be
> across
> several pages. In my library, I have a concept called "Data Access
> Objects". It makes development of the data layer very easy - almost
> mindless as a matter of fact - and I can actually create an entire tier
> that
> completely decouples database calls from application logic completely.
> This is something pear doesn't do and I think this is essential for
> webpages because a) they need to change all the time b) database code, php
> code and html code on the same page is messy c) this is how large
> enterprise systems need to be built.
>
> I think that should give you an explanation on why PHP still needs to
> develop. If not, then PHP should outright states its goals and intentions
> to everyone because people like myself who are waiting for things to move
> forward (because we have a lot of code invested into the language already)
> want to move forward with it. That just isn't happening from my point of
> view.
>
> To argue your point about performance, look at any emerging technology in
> the past. History has shown that coming up with the technical solution
> that
> works and solves people's problem is essential. Once something is in
> place,
> then we start looking at how to speed it up. But if we don't even get to
> the point of it working and solving people's problems, then we aren't
> going anyway.
>
> Regards,
> Ken
>
> ----- Original Message -----
> From: "Ilia A." <[EMAIL PROTECTED]>
> To: "Ken Egervari" <[EMAIL PROTECTED]>; "Richard Heyes"
> <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
> Sent: Friday, April 12, 2002 5:48 PM
> Subject: Re: [PHP-DEV] The PHP Platform
>
>
>> Ken,
>>
>> Many classes and for Java & .NET and even php very own PEAR are libraries
> of
>> bloat. They offer some functionality and in exchange take away tremendous
>> amount of extra resources. That is not to say all those classes and
> libraries
>> are written poorly, many are not, however no matter how good is a wrapper
> it
>> will always add slowdowns.
>> In php (at least 4.X) loading large classes to memory is VERY memory
> cosuming
>> and loading huge libraries will put a large strain on system with medium
> to
>> heavy use.
>> Most people do not need PEAR or other assistance libraries for most of
> their
>> code, especially in PHP where standard functions are VERY easy to use and
>> their is a function for almost every occasion :)
>> In many cases, like with database layers you can avoid class by simply
> using:
>>
>> $db_type = 'mysql_'; or $db_type('pg_'); etc...
>> and then calling php's database manipulation functions with $db_type
> prefix.
>>
>> So, in my opinion creating class libraries is counter productive in PHP
>> enviroment. It makes sence in C & C++ to some degree where to open a
> socket
>> you need to do a good deal of work, so a class which accepts a socket &
>> domain and returns open socket may be very useful. But in PHP where
>> everything is already done for you there is little need for that IMHO.
>>
>> On April 12, 2002 04:53 pm, Ken Egervari wrote:
>> > Hello Richard,
>> >
>> > I don't think people really understand me correctly. Pear is small in
>> > comparison to the Java Platform or the .NET Framework. My library
>> > extremephp.org is probably around 4-5 times bigger than PEAR and it's
> not
>> > even close to being finished yet. There could be much more to develop
> to
>> > make PHP an even greater language to use, but it's not keeping up.
>> >
>>
>> --
>> Ilia Alshanetsky
>> FUDforum Developer
>> http://fud.prohost.org/forum/
>>
--
PHP Development Mailing List <http://www.php.net/>
To unsubscribe, visit: http://www.php.net/unsub.php
