On Fri, Jul 15, 2016 at 5:48 PM, Jon Paris <jon.paris@xxxxxxxxxxxxxx> wrote:
PHP has succeeded in part because these many of these issues were addressed. Install via a familiar process, package install via the Zend console. Sure you occasionally have to dive down to get some things done but by the time you need that you have a fully functioning system up and running.
Now tell me that node.js is that easy? Or Python?
Aaron Bartell wrote, in response:
We're not comparing apples to apples. Node.js and Python, as languages,
~are~ as easy (actually easier) than PHP to install. It's when we get into
native language extensions that we get into discussions like this. PHP has
the same issue(s). It's not as noticeable with PHP because Zend has taken
the time to port some of the more popular native stuff.
I'm with Jon on this, actually. What Aaron says is *correct*. But it
doesn't matter, at least to a good chunk of the midrange community.
The bottom line is that the Zend PHP experience on the i has been
smoother and generally better, by typical midrange standards, than the
experience with Node.js, Python, and 5733-OPS in general.
PHP on the i has had the advantage of active involvement and support
from Zend, as well as a head start of several years vs. the latest
crop.
I have long thought IBM should have done a truly comprehensive Python
port. There have been Python articles on developerWorks for ages (e.g.
the "Charming Python" series by David Mertz, which dates back to
2001). I thought it would be a great fit, both because the language
itself feels (to me) like something that would be more accessible and
intuitive to midrangers than alternatives like Ruby; and because it
seemed like it was already a familiar and commonly used language in
some quarters of IBM itself.
It would have had the potential to be a superset of what Zend PHP for
i was, because Python has powerful and mature libraries for many more
things than PHP. And if one guy (Per Gummedal) can build something as
easy to install and easy to use as iSeriesPython, using ILE C, then
surely IBM could make a truly great version for PASE, or so I thought.
(Actually, what I was really hoping was that IBM take over and expand
Gummedal's project.)
I still think 5733-OPS has a chance. But it's trying a different
approach than Zend PHP for i, and it remains to be seen whether it
will turn out better.
John Y.
As an Amazon Associate we earn from qualifying purchases.