Hello Newsgroup,

Looks like PHP yanked the rug out from under us again, they've decided
to declare a cl*** DateTime and changed the way objects are cast into
strings.

5.1 -> 5.2 breaks anything along those lines.

You know, this wouldn't really bother me, I'd say the hell with PHP.

But.... clients demand PHP. They insist on it in fact.

If I try to explain that PHP is not good for anything more than simple
applications due to it's non-planning, they won't believe me. They'll point to
all these other applications that are written in PHP as "proof" that it's a
solid language.

I'm working with one of those applications right now, vtiger. I could probably
fix it, make it work with PHP 5.2, but.. when the PHP folks decide to change
the rules again, I'll be in hot water.

I /need/ this contract, moving on isn't an option. How do I convince someone
that PHP is not suitable for something of this scale when they see online demos
and so-forth that are obviously working?

Simply downgrading PHP isn't an option. Even if I did that, what will I do when
the earlier version is no longer available? for example, 5.1 would not work with
IMAP. (IMAP is needed by vtiger, but compiling the 5.1 version of PHP gave a bunch
of mime errors when using imap)

I don't have much of a chance at convincing him to go with something else, but
if a perl (or even java) package had the vtiger glitter it might (like, 1 in a
thousand chance..) be OK with him. He's one of these "more features are
better", even if the features are never actually used types of people. He doesn't
care one bit about the underlying design.

I'd like a CRM that's written really well with good, solid design principles that can
be extended. I /do/ care about underlying design.

I would really like to find some good solid arguments against PHP that a
non-programmer might actually listen to.

How would you talk someone out of PHP?

Jamie
--
http://www.geniegate.com Custom web programming
EMAIL REMOVED (rot13) User Management Solutions

Fleeing from the madness of the AT&T ASP.att.net jungle
Jamie <EMAIL REMOVED> stumbled into news:alt.www.webmaster
and said:

> Hello Newsgroup,

How do you do?

> Looks like PHP yanked the rug out from under us again, they've decided
> to declare a cl*** DateTime and changed the way objects are cast into
> strings.
>
> 5.1 -> 5.2 breaks anything along those lines.
>
> You know, this wouldn't really bother me, I'd say the hell with PHP.
>
> But.... clients demand PHP. They insist on it in fact.
>
> If I try to explain that PHP is not good for anything more than simple
> applications due to it's non-planning, they won't believe me. They'll
> point to
> all these other applications that are written in PHP as "proof" that
> it's a
> solid language.
>
> I'm working with one of those applications right now, vtiger. I could
> probably
> fix it, make it work with PHP 5.2, but.. when the PHP folks decide to
> change
> the rules again, I'll be in hot water.
>
> I /need/ this contract, moving on isn't an option. How do I convince
> someone
> that PHP is not suitable for something of this scale when they see
> online demos
> and so-forth that are obviously working?
>
> Simply downgrading PHP isn't an option. Even if I did that, what will I
> do when
> the earlier version is no longer available? for example, 5.1 would not
> work with
> IMAP. (IMAP is needed by vtiger, but compiling the 5.1 version of PHP
> gave a bunch
> of mime errors when using imap)
>
> I don't have much of a chance at convincing him to go with something
> else, but
> if a perl (or even java) package had the vtiger glitter it might (like,
> 1 in a
> thousand chance..) be OK with him. He's one of these "more features are
> better", even if the features are never actually used types of people.
> He doesn't
> care one bit about the underlying design.
>
> I'd like a CRM that's written really well with good, solid design
> principles that can
> be extended. I /do/ care about underlying design.
>
> I would really like to find some good solid arguments against PHP that a
> non-programmer might actually listen to.
>
> How would you talk someone out of PHP?

You can't ...

There are evangelists, enthusiasts, folk that feel comfy running with the
herd and twats that simply can't believe they backed the wrong horse -
give up on them. There's no hope here.

Then you have all the other techs that wouldn't touch PHP for anything
more complex than "Hello World" - no work to do here.

Ahh - then there's clients. Black box them - they really have no need to
know /how/ your latest ASAD app actually performs its magic. Unless the
client is the Project Manager paying by the hour, then you give them
exactly what they ask for regardless of the obvious fact that this is not
what they want. If you feel bad about thisn you can document your
concerns - but I'd caution against the round-robin "let me bring this to
your attention" type mail which won't do you (or the project) any favours.

PHP - Pure Hype Programming ?

--
William T***o

http://williamt***o.com/words/what-is-usenet.asp

And lo, Jamie didst speak in alt.www.webmaster:

> Hello Newsgroup,
>
> Looks like PHP yanked the rug out from under us again, they've decided
> to declare a cl*** DateTime and changed the way objects are cast into
> strings.
>
> 5.1 -> 5.2 breaks anything along those lines.

You think that's bad? PHP 6 "breaks" a lot more:
http://www.corephp.co.uk/archives/19...for-PHP-6.html

But do you want to know something interesting? None of these changes,
either in 5.2 or the proposed changes in 6.0 will affect any of my
scripts. Why? I keep up with best practise.

The developers of PHP know they have a badly planned language; they knew
years ago. They started on the long road of cleaning house some time ago
and thus these changes have been hinted at for just as long.

Best practise is even more important when coding in PHP in comparison with
HTML. PHP "standards" wait for no browser and are constantly evolving.
By all means, get creative, but when your programs start depending on
deprecated functions/syntax and vaguely defined methods such as how
objects are cast into strings, you're just asking for trouble.

> But.... clients demand PHP. They insist on it in fact.
>
> If I try to explain that PHP is not good for anything more than simple
> applications due to it's non-planning, they won't believe me. They'll
> point to
> all these other applications that are written in PHP as "proof" that
> it's a solid language.

It's a bumpy road in PHP-land at present, true. Reigning in an
out-of-control library of functions is not an easy task, and for certain
they're going to burn a few bridges of backwards compatibility. However,
it's only common sense that what the PHP team advocates as "best practise"
will be the last things they consider for changes.

> I'm working with one of those applications right now, vtiger. I could
> probably
> fix it, make it work with PHP 5.2, but.. when the PHP folks decide to
> change
> the rules again, I'll be in hot water.

This isn't your fault, technically. vtiger should have been planning for
changes such as these a long time ago.

Doesn't vtiger have a team to deal with these compatibility issues? Why
are you using a host that updates to the bleeding edge without asking the
client? If they are using a dedicated server, why not tell them to stick
with 5.1.x until vtiger releases an updated version?

> Simply downgrading PHP isn't an option. Even if I did that, what will I
> do when the earlier version is no longer available?

The PHP Group stores releases for a LONG time:
http://www.php.net/releases/index.php

You needn't worry.

> for example, 5.1 would not work with
> IMAP. (IMAP is needed by vtiger, but compiling the 5.1 version of PHP
> gave a bunch
> of mime errors when using imap)

How about downgrading to 5.0.x then? Or even 4.x? If you "need" this
client, I'm certain that fixing these little incompatibilities would be
much less time consuming than switching to a new language and all of the
rebuilding it would entail.

> I would really like to find some good solid arguments against PHP that a
> non-programmer might actually listen to.

How about: PHP is going through a rough spot right now as previous design
mistakes are corrected over an extended period of time?

> How would you talk someone out of PHP?

I wouldn't. Rather, I would quit whining and see about making the client
happy.

Grey

--
The technical axiom that nothing is impossible sinisterly implies the
pitfall corollary that nothing is ridiculous.
- http://www.greywyvern.com/orca#search - Orca Search: Full-featured
spider and site-search engine

EMAIL REMOVED (Jamie) wrote:

>Hello Newsgroup,
>
>Looks like PHP yanked the rug out from under us again, they've decided
>to declare a cl*** DateTime and changed the way objects are cast into
>strings.
>
>5.1 -> 5.2 breaks anything along those lines.
>
>You know, this wouldn't really bother me, I'd say the hell with PHP.
>
>But.... clients demand PHP. They insist on it in fact.
>
>If I try to explain that PHP is not good for anything more than simple
>applications due to it's non-planning, they won't believe me. They'll point to
>all these other applications that are written in PHP as "proof" that it's a
>solid language.
>
>I'm working with one of those applications right now, vtiger. I could probably
>fix it, make it work with PHP 5.2, but.. when the PHP folks decide to change
>the rules again, I'll be in hot water.
>
>I /need/ this contract, moving on isn't an option. How do I convince someone
>that PHP is not suitable for something of this scale when they see online demos
>and so-forth that are obviously working?
>
>Simply downgrading PHP isn't an option. Even if I did that, what will I do when
>the earlier version is no longer available? for example, 5.1 would not work with
>IMAP. (IMAP is needed by vtiger, but compiling the 5.1 version of PHP gave a bunch
>of mime errors when using imap)
>
>I don't have much of a chance at convincing him to go with something else, but
>if a perl (or even java) package had the vtiger glitter it might (like, 1 in a
>thousand chance..) be OK with him. He's one of these "more features are
>better", even if the features are never actually used types of people. He doesn't
>care one bit about the underlying design.
>
>I'd like a CRM that's written really well with good, solid design principles that can
>be extended. I /do/ care about underlying design.
>
>I would really like to find some good solid arguments against PHP that a
>non-programmer might actually listen to.
>
>How would you talk someone out of PHP?
>
>Jamie

You're basically screwed. Non-programmers that think they know all
about it are impossible to deal with. The reason you're screwed is:

>I /need/ this contract, moving on isn't an option.

This is not to say that PHP cannot be used for large applications,
it's just a major pain. I recently moved to a different shared server
running a newer version of PHP and they had made a bunch of spurious
changes, like fread() bellyaches if filesize() is zero in the newer
release when there's no reason for it. Incompatibilities in new
versions is typical of ad-hoc languages, as are inconsistencies in
syntax and function naming, for example isset() vs is_array() don't
follow the same naming convention, and for krissakes when you have to
unset() a global variable the syntax is godawful. The problem is that
there are so many "programmers" out there who have never worked with
anything besides ad-hoc languages that they don't even realize that it
hurts.

The only thing I can suggest is to charge them enough money for
working in PHP to make it profitable. That's an argument even the
most brilliant non-programmer can understand. "Sure, be glad to write
it in PHP, cost you lots more but if you want it in PHP, you bet we'll
be glad to do that, send fifty grand in advance and we'll get right on
it."

--
Legacy browsers never heard of emerging standards.
(contact via http://www.ren-prod-inc.com/hug_soft)

EMAIL REMOVED (Jamie) wrote:

> How would you talk someone out of PHP?

Impossible. I stopped doing PHP work because the language was extremely
bad designed, and still needs a lot of work to come close to other
languages I use. It's crazy how someone can come up with a piece of junk
and the whole world thinks it's good, but that seems to happen more often
then I want to know :-)

So decide: do you really want to do PHP work or not. If it's a yes, you
have to learn with the fact that a part of your time, maybe even the
majority, will consist of rewriting code, and learning all
incompatibilities of versions, and fixing crap others have written.

If you don't, check out which language you want to use instead, or find
someone you can hire to do the PHP work.

--
John Need help with SEO? Get started with a SEO report of your site:

--> http://johnbokma.com/websitedesign/seo-expert-help.html

Jamie wrote:
> Hello Newsgroup,
>
> Looks like PHP yanked the rug out from under us again, they've decided
> to declare a cl*** DateTime and changed the way objects are cast into
> strings.
>
> 5.1 -> 5.2 breaks anything along those lines.
>
> You know, this wouldn't really bother me, I'd say the hell with PHP.
>
> But.... clients demand PHP. They insist on it in fact.
>
> If I try to explain that PHP is not good for anything more than simple
> applications due to it's non-planning, they won't believe me. They'll point to
> all these other applications that are written in PHP as "proof" that it's a
> solid language.
>
> I'm working with one of those applications right now, vtiger. I could probably
> fix it, make it work with PHP 5.2, but.. when the PHP folks decide to change
> the rules again, I'll be in hot water.
>
> I /need/ this contract, moving on isn't an option. How do I convince someone
> that PHP is not suitable for something of this scale when they see online demos
> and so-forth that are obviously working?
>
> Simply downgrading PHP isn't an option. Even if I did that, what will I do when
> the earlier version is no longer available? for example, 5.1 would not work with
> IMAP. (IMAP is needed by vtiger, but compiling the 5.1 version of PHP gave a bunch
> of mime errors when using imap)
>
> I don't have much of a chance at convincing him to go with something else, but
> if a perl (or even java) package had the vtiger glitter it might (like, 1 in a
> thousand chance..) be OK with him. He's one of these "more features are
> better", even if the features are never actually used types of people. He doesn't
> care one bit about the underlying design.
>
> I'd like a CRM that's written really well with good, solid design principles that can
> be extended. I /do/ care about underlying design.
>
> I would really like to find some good solid arguments against PHP that a
> non-programmer might actually listen to.
>
> How would you talk someone out of PHP?
>
> Jamie

I actually rather like PHP. That said, I don't consider PHP 5.x to be
usable. I'm sticking with 4.x until I see a good reason to switch to 6.x.

I see it as being much like with Apache. 1.3.x is very usable, and is
pretty much an industry standard, especially for commercial web hosting.
2.x has been out for a long time now, and is what Apache recommends
using, but there's a very good reason why the world isn't listening: 2.x
breaks everything you did for 1.3.x. And, it's a moving target, though
arguably not nearly as much as PHP 5.x has become.

4.x may be conceptually "broken," but it works quite well, and has many
supporting add-ons (like Smarty and PEAR) that make it superior (IMHO)
to the competition. 5.x may be "better," but as you've seen yourself,
even incremental upgrades make sweeping changes that break what works. I
think this may be why, much like with Apache, most commercial hosting
providers tend to stick with the older PHP 4.x -- simply because it is a
known element.

Out of curiosity, given that you are looking for an argument against
PHP, do you have an alternative language you prefer, such as ASP, JSP,
Perl, etc.?

--
David J. Hennessy
http://maidix.com/
http://choicesolver.com/
http://davidhennessy.net/

On Mon, 18 Dec 2006 11:36:04 -0500, GreyWyvern <EMAIL REMOVED>
wrote:

>But do you want to know something interesting? None of these changes,
>either in 5.2 or the proposed changes in 6.0 will affect any of my
>scripts. Why? I keep up with best practise.

By using Perl instead <bg>

</gets coat>

Matt

In <oyAhh.1431$Pq4.747@trndny08>,
"David J. Hennessy" <EMAIL REMOVED> mentions:
>I actually rather like PHP. That said, I don't consider PHP 5.x to be
>usable. I'm sticking with 4.x until I see a good reason to switch to 6.x.

I "upgraded" *my* stuff to PHP 5.0 as soon as I saw "At last! exception handling!"

Then.. watched the complaints come tumbling in. Basically had to re-write the
whole thing back to PHP 4 when I realized how many people were using it.

>I see it as being much like with Apache. 1.3.x is very usable, and is
>pretty much an industry standard, especially for commercial web hosting.
>2.x has been out for a long time now, and is what Apache recommends
>using, but there's a very good reason why the world isn't listening: 2.x
>breaks everything you did for 1.3.x. And, it's a moving target, though
>arguably not nearly as much as PHP 5.x has become.

I use apache 2.n, I like it, but.. I sort of gave up on mod_perl or anything
that is really tied in with the web server. I like the FastCGI / Tomcat approach
and tend to believe the web server should be a web server, not an everything-
under-the-sun server. (linux kernel seems to have the same problem, IMO maybe
I'm getting old LOL)

Incidently, I tried to compile an older version of apache for DragonFlyBSD, was
a no-go as it's configure script didn't recognize the platform. (first thought
was to completely isolate this PHP none-sense into a BSD jail so I had absolute
control over the libraries for testing, I might take another stab at that approach,
if I can. I quite like the ideas of DragonFlyBSD, it's simple and no-frills.

>4.x may be conceptually "broken," but it works quite well, and has many
>supporting add-ons (like Smarty and PEAR) that make it superior (IMHO)
>to the competition. 5.x may be "better," but as you've seen yourself,
>even incremental upgrades make sweeping changes that break what works. I
>think this may be why, much like with Apache, most commercial hosting
>providers tend to stick with the older PHP 4.x -- simply because it is a
>known element.
>
>Out of curiosity, given that you are looking for an argument against
>PHP, do you have an alternative language you prefer, such as ASP, JSP,
>Perl, etc.?

Perl or Java. (Been a few years since I've worked with java, it's not "fun" by
any means, however, Java is probably the better tool for the job) I'd prefer
to work in perl though.

Unfortunately, getting the client to see that, (well, vtiger in this case) isn't
going to work. Unless I knew of a Perl or Java CRM solution that has all the
fluff of vtiger. (and can be copied freely) As well as could some how convince
him to actually /look at/ such a solution.. it's a lost cause.

It's the glitter of a dozen or so sub-projects and add-ons that have sold him
on vtiger. Too bad the foundations aren't as visible as the interface. Too their
credit, vtiger itself was written pretty good, wish it used a "real" MVC, and lost
the "Smarty" templates, but I've seen much worse. (I know, some people like Smarty,
I don't in this case, just one more thing to go wrong)

Jamie
--
http://www.geniegate.com Custom web programming
EMAIL REMOVED (rot13) User Management Solutions

EMAIL REMOVED (Jamie) wrote:

>How do you walk into something, knowing it's going to blow up, and shield yourself
>from the fallout later? (not just in legal terms, but in relationship/reputation
>terms?)

You don't. You tell him you won't do it because you feel it's a bad
solution. You lose the client, probably. But you retain some
semblance of control and ethical responsibility. If you can't afford
to lose the client, charge him an arm and a leg because you'll be
earning it.

--
Legacy browsers never heard of emerging standards.
(contact via http://www.ren-prod-inc.com/hug_soft)

Jamie wrote:
> But.... clients demand PHP. They insist on it in fact.

Then give them the .php file extensions they expect to see...
but call functions in a language that really works.