PHP's declining(?) popularity

https://jaxenter.com/php-tiobe-sept-2019-162096.html
I think this is one of those things we get from voting no...

I might be wrong anyways :-?

If those specific rankings are legitimately a cause for concern then it would make sense to do some objective analysis to determine why the languages that are growing marketshare are growing.

From the list it seems Python stands out as having the most growth as a percentage.

Googling for "why has python become so popular" I get these articles.

I have copied their top level points but also included the text for the one point they all seem to have in common, that Python is simple, easy to use and easy to learn for new users (emphasis mine in all cases):

Want an easy, one-line improvement that costs the community nothing?
Python is primarily a CLI tool. (Yes, it has GUI and, more recently,
web bindings, but the vast majority of Python code I run into is
CLI-oriented.)

Why? Because those people who use Python don't realize that PHP is also
a CLI tool and an excellent one at that. Here's the website description
for PHP:

"PHP is a popular general-purpose scripting language that is especially
suited to web development.

Fast, flexible and pragmatic, PHP powers everything from your blog to
the most popular websites in the world."

There's nothing in those two sentences that readily indicates anything
other than "tool for website development." So people are going to look
elsewhere for a language that does CLI and therefore that's what they do.

I tried to get the following change (or something like it) added years
ago and it was initially approved but was later ignored or forgotten:

"PHP is a popular general-purpose scripting language that is especially
suited to web development and command-line scripting.

Fast, flexible and pragmatic, PHP powers everything from your blog to
the most popular websites in the world."

That's just one minor little text change on the homepage to make people
aware that PHP is more than just a web development scripting language
tool. I write all of my cron jobs in PHP. I write web scrapers in PHP.
I write servers in PHP that run as system services as root (uh...).
Pretty much everything I do is in PHP...but very little of it being
actual web development. Whenever I bring those things up to people, I
get funny looks that say, "PHP does CLI? Huh. Never knew that."

--
Thomas Hruska
CubicleSoft President

I've got great, time saving software that you will find useful.

http://cubiclesoft.com/

And once you find my software useful:

http://cubiclesoft.com/donate/

On Sun, Sep 15, 2019 at 6:33 AM Mike Schinkel mikeschinkel@gmail.com
wrote:

On Sep 14, 2019, at 5:18 PM, Olumide Samson oludonsexy@gmail.com
wrote:

https://jaxenter.com/php-tiobe-sept-2019-162096.html
I think this is one of those things we get from voting no...

I might be wrong anyways :-?

First of all, Olumide, this is in fact wrong, although the general topic
(language popularity and the reasons to it) is definitely worthy of
discussion.

The reason it's wrong is that TIOBE is a meaningless 'index' with a
methodology that's not only questionable - but is rather downright
idiotic. It's not just off or inaccurate - it's practically a random
number generator.
See for yourself:
https://www.tiobe.com/tiobe-index/programming-languages-definition/

The RedMonk Language Rankings has a much more reasonable methodology, is a
lot more stable, and there, PHP is repeatedly at the top 5 languages and
not losing any steam in both absolute and comparative measures:
https://redmonk.com/sogrady/2019/07/18/language-rankings-6-19/

If those specific rankings are legitimately a cause for concern then it
would make sense to do some objective analysis to determine why the
languages that are growing marketshare are growing.

Mike - even though specifically the TIOBE index isn't a cause for virtually
anything, the rest of your analysis is still relevant - as the key takeaway
you're basing it on - Python's growth - is also reflected in RedMonk
rankings.

Thomas - I also wholeheartedly agree with your suggestion. That's why we
worked on FFI - to open the door for PHP to enter new areas. Even JIT is,
for the most part, not really relevant to the common Web case and would be
a lot more impactful in other types of workloads. And there may be other
things we can do. But you're right - if we don't find a way to position it
for these use cases in people's minds - it won't move the needle.

Zeev

I'm not going to go into my opinions on rankings - others have already
done enough of that. I concur with them - most rankings are meaningless.
PHP is also not, in my opinion, dead or dying. It doesn't enjoy the
monopoly it once had on web-based development, but it's also still very
popular.

PHP has been my primary language for a long time. I've recently been
undertaking a small personal project in Python - namely because there's
no well maintained PHP equivalent of the library I want to use
(gmusicapi - Google Play Music has no official documented API).

In my opinion, Python is popular on the command-line / linux partly
because it's become a replacement for Perl. Perl used to be king for
doing a lot of scripting and anything remotely complex, then they went
off on their Perl 6 voyage and now I don't think anyone knows what's
going on with it or has the time to care.

Python also has a reputation for being a language for doing "big data" /
"machine learning" / "AI" things, and since that's a big thing, it gets
picked a lot for those reasons.

It's also used in places like Plex and XBMC/Kodi plugins where I think a
number of people get their first "I want to program" itch. Wordpress
(plugins) used to be a big draw for PHP here, but I think with the rise
of alternatives (static site builders, use of other languages, increased
capabilities on browsers) this has waned over the years.

Maybe it's coming from PHP and I was specifically trying to avoid having
to deal with writing / using a Python HTTP server and getting a setup
similar to PHP-FPM, but in my opinion setting up anything web-based in
Python is kind of a mess - there's a lot of alternatives and no clear
easy option. For me, PHP, with the simplicity of PHP-FPM is, in my
opinion, clearly better here.

One area where I think Python really stands out compared to PHP tho is
its "new user experience". Starting with the official tutorial (
https://docs.python.org/3/tutorial/ ): While I was obviously coming in
with a lot of programming knowledge, I felt Python's tutorial is very
well done. Its well laid out index means you can quickly skip around
when you already know what you want to achieve (but not how to achieve it).

Compare this to the official PHP tutorial ( https://www.php.net/tutorial
) which to me feels pretty much only covers a basic HTML form. The
"What's Next" section directs you to http://talks.php.net/ which has
very limited content from the past 5 years and a lot from before that.
It also has a theme that hurts my eyes to look at (kind of ironic that
the "PHP Presentations" site has horrible presentation =P ). The "What
Can PHP Do" link takes you to a page that is pretty much a wall of text.
It spends a lot of time talking about things (eg. listing supported
operating systems) that probably 99% of users don't care about at this
point as well as linking to effectively dead projects like PHP-GTK.

If you then try to proceed onwards, one of the next things people
frequently want to do is storing information in Sessions (
https://www.php.net/manual/en/book.session.php ) (from my experience
helping users in the likes of Freenode's ##php ). While I understand
this section as a long time PHP developer and it does cover a lot of
technical information, if you're coming into this as a newbie developer
while you can sort of get some basic session code going on, many newer
developers do hit / have questions on slightly more complex topics like
session locks or session file management and are, I believe, quickly
left with their head spinning.

Python also has a user editable wiki with the Beginners Guide (
https://wiki.python.org/moin/BeginnersGuide ) being practically the
first link after the main documentation links on the Docs section of the
Python site ( https://www.python.org/doc/ )

I think the entire "new user experience" of PHP could do with review.
Some suggestions to start:

• Link directly to the tutorial from the front page of php.net
• Look at removing the https://www.php.net/docs.php page and replacing
  it by linking directly to the manual.
    * PHP 4 documentation, "More documentation" and downloads could be
  linked as an / from an appendix.
    * There's already a language switch on the top right of the page for
  swithcing to other languages
• Move contributors to an appendix - it unnecessarily takes up a
  significant amount of space at the top of the manual index forcing users
  to scroll down to the content they want. We're all very grateful to the
  manual contributors, but it's not what 99% of users are there for.
• Link directly to Language Reference (
  https://www.php.net/manual/en/langref.php ) and Features (
  https://www.php.net/manual/en/features.php ) from the tutorial

talks.php.net:

• Update links to use HTTPS (minor, but the "Not secure" is still ugly)

• Link from the front page of php.net

• Make it clear and easy for users to submit new talks and otherwise
  contribute to this section

• Improved theme that doesn't make me want to scratch my eyes out. Avoid
  unnecessary animations (the elephpant is cute but it keeps pulling my
  brain from the actual content). (I might look into submitting some
  simple changes if/when I can find the repo for this)

• Review and tidy up older talks
    * Remove anything that's completely out-of-date / doesn't work in PHP 7.x
    * Where multiple talks cover the same topic / material, pick the best
  / more up todate ones (for example there's like 3 dozen XDebug talks)
    * Some talks are broken - for example
  http://talks.php.net/show/vxml_php_2004 gives an error when trying to go
  to the next slide (bug filed: https://bugs.php.net/bug.php?id=78542 )

• Consider a user-editable wiki
    * Talks would go well here, I think
    * See examples where other OSS projects do this - Python, Arch &
  Gentoo Linux

A wiki might not be necessary if/when users can submit changes to the
above mentioned content via GitHub assuming it's in an easy-to-edit
format, which I know is being worked on. I understand the challenges of
moderating wiki's (I was an moderator on the unofficial Gentoo Wiki for
a time) and a PR based system does have the advantage of completely
preventing spam / bad faith edits making it to the live documentation/site.

A wiki would however provide a place where unofficial links can be made

• for example I feel it kind of weird that Composer / Packagist gets no
  mention on php.net. I understand it's a separate project and php.net
  doesn't want to prevent or dissuade competition, but it is also the
  defacto package manager for PHP and an easy way to find libraries for
  getting things done.

With regards to academia, we were taught PHP in the web module on my
Computer Science course at uni (England), but as can be the case the
material was out of date and hadn't been revised in a few years. I have
no idea how to solve this problem because it stems from teachers who
don't care about the subject.

AllenJB

2019.09.15 06:32 Mike Schinkel rašė:

https://medium.com/@trungluongquang/why-python-is-popular-despite-being-super-slow-83a8320412a9

1. End-users just don’t care (about slower performance)

Which means that your code is running data processing and not interactive
user facing frontend. Put 10 second delay in your web page and ask
viewers, if they care.

If you claim that language is good or popular while being slow, you are
not telling where this language is used. Super slow interactive frontends
are not acceptable even if it is reality in corporate world. Designers of
such slow frontends are thinking about features that they have to put in
there and not about user experience or how to implement those features
without making end user drink five coffee cups to fill one form.

--
Tomas

The lack of concurrent PHP features in 2019 is, in my opinion, sufficient
reason not to use it for CLI projects.
Some companies do not welcome installing extensions like Swoole, the
language should have native concurrent features.

Daniel

El dom., 15 sept. 2019 a las 0:33, Mike Schinkel (mikeschinkel@gmail.com)
escribió:

On Sep 14, 2019, at 5:18 PM, Olumide Samson oludonsexy@gmail.com
wrote:

https://jaxenter.com/php-tiobe-sept-2019-162096.html
I think this is one of those things we get from voting no...

I might be wrong anyways :-?

If those specific rankings are legitimately a cause for concern then it
would make sense to do some objective analysis to determine why the
languages that are growing marketshare are growing.

From the list it seems Python stands out as having the most growth as a
percentage.

Googling for "why has python become so popular" I get these articles.

I have copied their top level points but also included the text for the
one point they all seem to have in common, that Python is simple, easy to
use and easy to learn for new users (emphasis mine in all cases):

https://medium.com/@trungluongquang/why-python-is-popular-despite-being-super-slow-83a8320412a9
<
https://medium.com/@trungluongquang/why-python-is-popular-despite-being-super-slow-83a8320412a9

1. End-users just don’t care (about slower performance)
2. More Productive
   "First and foremost reason why Python is much popular because it is highly
   productive as compared to other programming languages like C++ and Java. It
   is much more concise and expressive language and requires less time,
   effort, and lines of code to perform the same operations. The Python
   features like one-liners and dynamic type system allow developers to write
   very fewer lines of code for tasks that require more lines of code in other
   languages. This makes Python very easy-to-learn programming language even
   for beginners and newbies. For instance, Python programs are slower than
   Java, but they also take very less time to develop, as Python codes are 3
   to 5 times shorter than Java codes. Python is also very famous for its
   simple programming syntax, code readability and English-like commands that
   make coding in Python lot easier and efficient."
3. Execution Speed does not matter as much as Business Speed

https://www.kdnuggets.com/2017/07/6-reasons-python-suddenly-super-popular.html
<
https://www.kdnuggets.com/2017/07/6-reasons-python-suddenly-super-popular.html

1. Python Has a Healthy, Active and Supportive Community
2. Python Has Some Great Corporate Sponsors
3. Python Has Big Data
4. Python Has Amazing Libraries
5. Python Is Reliable and Efficient
6. Python Is Accessible
   "For newcomers and beginners, Python is incredibly easy to learn and use.
   In fact, it’s one of the most accessible programming languages available.
   Part of the reason is the simplified syntax with an emphasis on natural
   language. But it’s also because you can write Python code and execute it
   much faster."

https://www.techrepublic.com/article/why-python-is-so-popular-with-developers-3-reasons-the-language-has-exploded/
<
https://www.techrepublic.com/article/why-python-is-so-popular-with-developers-3-reasons-the-language-has-exploded/

1. Ease of learning
   "Python is the closest language to what I call 'an instant gratification
   language,' meaning with very little code, it can accomplish so much, even
   if you are a novice programmer," said Karen Panetta, an IEEE Fellow and
   Dean of Graduate Engineering at Tufts University. "This is because Python
   reads like English, which makes it more conducive for a broad level of user
   audiences to learn. Many of the 'nerdy' low-level details we used to worry
   about in other languages, such as declaring types of variables or
   arguments, are handled by Python, so it makes programming very flexible and
   easier to use than other languages."
2. The explosion of AI, machine learning, and data science in the
   enterprise
3. A large developer community

https://www.skillsoft.com/blog/2018/06/top-5-reasons-why-python-is-so-popular/
<
https://www.skillsoft.com/blog/2018/06/top-5-reasons-why-python-is-so-popular/

1. Python is efficient
2. Python has an active community
3. Python is simple
   "With a shorter learning curve than other languages, say Java or C++, and
   understandable and readable syntax, you don’t need to be a programmer to
   start applying Python to everyday tasks. Python automatically takes care of
   things like garbage collection and even closes files, opened via the ‘with’
   statement, for you. People starting out may also find the use of
   indentation to signify the start and end of loops, functions, classes and
   code blocks easier than tracking down the traditional opening and closing
   curly braces."
4. Python is in academia
5. Python is on trend

https://eplexity.com/6-reasons-the-python-programming-language-is-so-popular/
<
https://eplexity.com/6-reasons-the-python-programming-language-is-so-popular/

1. Ease of use
   "Since its creation in the late 1980s by Guido van Rossum, Python has been
   specifically designed to be a general-purpose language. The simplicity of
   Python, and its easily human-readable syntax are two reasons why the
   language is so popular among seasoned coding professionals and Computer
   Science 101 students alike. Python is also an interpreted language, which
   means that you can quickly experiment with changes to the code base."
2. Supportive community
3. Corporate sponsors
4. Libraries and frameworks
5. Use in big data and machine learning
6. Efficiency

Notice in none of these articles is there any mention of static typing or
strictness or correctness as an attribute that the authors think
contributes to why Python is gaining success.

So my takeaway would be that if we wanted to see PHP start gaining
marketshare again we should focus on features that make it easier to use
and spend less time on trying to raise the bar for the skill level it will
take to program in PHP8.

#jmtcw

-Mike

P.S. Other ways to grow market share could be to:

1. Strive to support data science in a major way

2. Run on Single Board Computers like Raspberry Pi

3. Run a CLI via a single executable (i.e. package PHP runtime and PHAR
   files together so PHP scripts could be distributed without requiring a
   correctly installed version of PHP)

4. Empower developers to create cross-platform desktop, also contained
   with a single executable

5. Empower developers to create cross-platform mobile apps

6. Pursue corporate sponsors (too bad the Facebook ship already sailed.)

7. Find ways to get PHP taught in academia

8. Put more effort into driving adoption for PSRs like 6,7, 11, 14, 15,
   16, 18 and in identifying and releasing more.

9. Strive to make PHP the best solution for AWS Lamba and other serverless
   solutions

10. Embrace all of userland PHP by giving them a voice in the direction of
    PHP, instead of limiting that voice to only those currently with a vote via
    phpinternals.

Of course just because I included it in the list above does not mean the
PHP team needs to pursue (any of) them. I am just pointing out areas where
I think PHP could see some growth if PHP pursued it.

Also, any of the above would require the PHP internals community to join
together for the common-good rather than taking no-comprised positions that
each seek one person's view of PHP perfection while blocking everything
else.

IMO anyway.