Re: [PHP4BETA] Re: [PHP-DEV] PHP 4.0.0Release packaged

Hello Ron,

On 24-May-00 05:55:18, you wrote:

>> >Well, what's wrong with your website? I _still_ havn't gotten a good
>> >download of metabase because the damn thing won't resend my password,
>> >(which I didn't get to choose, which screwed up my login). I can't read
>> >the metabase
>> Passwords are sent automatically by e-mail. Has it crossed your mind that
>> your server may be bouncing messages to you sent by my site?

>Yup. If your messages are poorly formed, it may set off spam filters....

I don't see anything poorly formed in my messages.  What I see is that many
spam filters are fallible, not to speak that there is a chance that
somebody that doesn't like your face to get even by reporting your site to
mail abuse sites.


>> Since I can't get through to you over, I hope the list excuse me for using
>> this message to inform that there is a recent account with your name and
>> address that says that your password is the one that was sent in the
>> e-mail.

>ARGGH! I never got a good password to check this with! :-)

That's odd because I can't seem to found any bounces from messages sent to
you. I'll make your password be sent again.


>> There is another older account with a name similar to yours that says that
>> the password was reset by the system on 2000-05-19.  Either way, if you
>> didn't get a password, it's certainly an e-mail problem that transcends the
>> best of my efforts.

>Shall I just keep re-registering? Is your site SPAM filtered on any of
>the blacklists?

I haven't found any SPAM list that could be checked from the Web and not by
mailing some person.  Anyway, my site is hosted by a ISP as a virtual
server.  Other ISP clients may have got the machine address into a blacklist
and I'll never know about it.


>> >documentation because your website is so screwed up that the documentation
>> >can't be found without your broken user/pass being fixed. I can't get
>> That's not broken, that's meant to be for justified reasons.
>> >reliable information, I can't get a decent download, I send you the
>> >information to get it fixed, and YOU IGNORED IT. Why can't you act like
>> >the 
>> I don't recall having received any message in those terms.  When have you
>> sent it?

>About three weeks ago, I think.

Now I know that I found your slight reference to the problem deep in a
message about a completely unrelated subject that was queued in my 120+
personal mail box with low priority.  You should have mailed me a separate
message regarding the subject.


>> I even try ask those people to show what they feel in
>> public, but I understand that they just don't feel like being beat up as
>> well like I am having to put up.

>Well, you see mine. :-)

I am talking of people that support my views as they are.


>> >Well, you've had _mine_... you write a lot, and different people support
>> >different things. I want a unified database abstraction layer, but I think
>> >the idea of making PHP more like closed source is totally wrong. Closed
>> Never meant it that way.

>Well, quality standards more like closed source? Many quality standards
>cause problems in software projects. They make the product later, less
>performant, more marginalized.

Not if they are done right.  You need to read books like "Writing Solid
Code" to understand why Microsoft software is so buggy, although they used
to be worse.

Anyway, systematic quality assurance procedures may pay a lot more, than
the "let's have a go to see if looks like it is working" approach.  For
instance, Metabase is getting many new drivers to support different
databases.  Metabase drivers have to support so many features that the
"let's have a go" approach wouldn't guarantee any quality of the drivers
regardless of the qualifications of the developers that wrote them.

So, I developed a test suite script that tries the features and verifies
if the results are what is expected.  With this I found not only many
little bugs that were hard to notice, but also the PostgreSQL problem with
PHP 4 that I mentioned and also a bug in MySQL 3.23 that gives you problems
with auto_increment fields that start at negative values.  Anyway, now I am
more confident of the quality of the Metabase drivers that exist and those
that are about to come.

Of course not all developers that contribute to PHP are aware of test
engineering procedures and some times that may be a problem, but this is
one of the systematic quality assurance procedures that pays to be
implemented.


>> > than open source. It
>> >takes longer to get bug fixes, because they are tied up in Quality
>> >Assurance for two months, not two weeks.... 

>This is obvious.

>Here are some other problems to consider:
>Freezing code means that bugs are not fixed, but _frozen_. Most final

Wrong. Code freeze, means no new feature code is added.


>releases ship with frozen bugs for this reason, it's too late to unfreeze,
>change, and start up the freeze periods again.

That's not the problem.  The problem is that marketing people put pressures
to release sooner rather than later.


>> >in your criticisms, you often
>> >bring out support from different people, depending on their concerns. 
>> So, you feel I am wrong proposing measures to assure better software
>> quality in PHP?

>I think that quality measures can make a product better, and also worse.
>it takes careful thinking and planning to make it better, it takes
>arguments, and disagreements, to come to a good solution to make a
>product better and still let it ship.

It's nothing like that.  Most discussions take on product specifications
process.  Programs like PHP lack of that because code is added on good will
basis.  Code is usually added ad-hoc.  You can't expect much better from
Open Source model.

Decisions on new features should be done on early stages of development.
Procedures to test those features should also be planned, but many
contributors have no idea what test engineering is and why it is so
important for quality assurance.


>> >> >everybody here respects the endless, unpaid hours of work that were put
>> >> >into PHP and into php.net
>> >Go team, go!
>> There it is the noisy crowd cheering for Golias! :-)

>I cheer for Goliath _and_ David. I would say you are a Goliath... the PHP
>team knows you, and you're certainly not meek and quiet like David. :-)

I don't agree, I am the one that has to fight against arrogance here.


>> >It has to be constructive. What was constuctive in what I said?
>> >a) Don't force registration on your website.
>> If only you knew that is one vital thing that lead to the site success.
>> Let me explain, people subscribe to be able to download.  All downloads are
>> accounted rigourously.  With those figures I can compute top charts that
>> are displayed in the site.

>You can do this *without* requiring a username and password to be mailed.

How do you assure for the rigor of the download counts. Without passwords
one crazy author may just download his package a lot of times to make it
raise in the charts.


>> Conclusion, since I added the top charts page the number uploaded classes
>> boosted.  Without requiring people to register it would be possible to
>> compute accurate and credible charts.

>But less people will download. You said yourself that you are getting mail

Less than what? I know that some won't subscribe because they don't want to
be identified when they download.  Too bad, I'd rather live without their
downloads than without the ability to stimulate authors to upload more and
better code to have their names figure in credible charts!


>bounces. Do you read why they are bouncing?

Many reasons.  Some have their mailboxes full.  Others just redirect
messages to non existing mail boxes.  Others have broken auto-response
deamons.

Anyway with 8000+ subscribers I can't be bothered to check all the bounces.
I just suspend message delivery to them although they are not unsubscribed
because of that.



>> >> In
>> >> the real world people are paid to judge and criticise.  Many of those
>> >> are called consultants. I guess I get some hostile feedback to my
>> >> criticisms because I am not charging for it. Maybe I should. :-)
>> >You should. It pays well....
>> Believe me that I tried.  I worked for ISP that develops web sites for
>> clients. They use ASP because that's all that they know. They fear that
>> they may get lost in the change.  They know ASP/MSSQL/IIS/NT works for
>> them.  They don't know that for PHP.  PHP is not advertised in the real
>> world.

>I have one client who _refuses_ to use ASP, because PHP is what they know.
>It's not about PHP, or ASP, it's about people fearing changes.

Right, but many people would change to PHP, first if they were aware that
it is much better and second because can trust that it is a step forward,
not backwards.  Scared people give higher relevance to any bug or lack in
the documentation that they may find.


>> Take a look at Web development magazines and you see that the difference
>> articles on ASP are more or less about 10 to 1 on PHP.

>Yes, it takes as much as ten times the work. :-) 

No pain, no gain.


>> The low presence of
>> PHP in the media that many Web development companies use to get informed
>> make them believe that ASP is what everybody is using.

>Very true. So shout it from the mountaintops! Don't mask your pages in .html,
>proudly display .php, .php3, .phpeatsASPforbreakfast!

I don't think so.  I am supporter not a fanatic.  Fanatics are often an
embarrassment to themselves because of their obcessive and often hostile
attitudes against those that don't support the same. Fanatics often don't
notice how counterproductive that may be and that usually non-fanatics make
people stay away to not be associated with them.

Another point, is that I follow the convention to use an extension for my
pages that is recognizable by browsers as being of the actual data type for
when the pages are saved to disk to be read on later.



>> They only know
>> what they see in the media.  It's pointless to laugh at their ignorance
>> because this lack of media recognition affects all of us big time.

>So we have to make it.

Somebody has.


>> To summarize, I tried to be helpful and bring up this feedback from the
>> real world.  It's a shame how some people don't bother to stop and try to
>> understand what I am saying instead of just killing the messanger!  :-(

>It's not just the messenger, Manuel, it's how the messenger is delivering
>the message. 

I may sing and dance while I tell you how unstatisfactory things are.
Still that doesn't change the message itself.


Regards,
Manuel Lemos

Web Programming Components using PHP Classes.
Look at: http://phpclasses.UpperDesign.com/[email protected]
--
E-mail: [email protected]
URL: http://www.mlemos.e-na.net/
PGP key: http://www.mlemos.e-na.net/ManuelLemos.pgp
--