Python on Pick

[Peter McMurray]

I was bemused when I saw this language being touted as an improvement by Rocket. My initial feeling was WHY?

However I was in the local library yesterday, it is housed in the primary school, and I saw books on coding for primary students.

I know that some ministers have been saying that children from Grade 1 on should learn coding. My thoughts on that were English, a second language, Maths and Accounting should all precede that.

Then I saw these book. Typical children's books say 30 pages long and lots of pictures. They covered topics from loading Python through to writing interactive screens for the Web using Loops, Lists and checking answers etc..

What is it about Python that is so fascinating?

[Bob Dubery]

For schools it's good. It's free, for a start, so there is no problem with having to negotiate a license, have that audited and so on.

And there's object orientation, so young programmers get used to that way of working. 

[George Gallen]

My biggest gripe with Python is the blocks are indent oriented, not {}

and the indents have to be the same spacing - so sometimes it's difficult to copy code

and the indents have to be the same character (tab or spaces) - so if one program uses tab for indents and others use spaces

      and you copy code - it will "look" fine, but won't work.

But.....Python is the preferred language of Raspberry PI programming - which can also might entice younger blood

to programming.

George

---

From: mvd...@googlegroups.com <mvd...@googlegroups.com> on behalf of Bob Dubery <mega...@gmail.com>
Sent: Sunday, July 30, 2017 10:50 PM
To: Pick and MultiValue Databases
Subject: [mvdbms] Re: Python on Pick

 

--
You received this message because you are subscribed to
the "Pick and MultiValue Databases" group.
To post, email to: mvd...@googlegroups.com
To unsubscribe, email to: mvdbms+un...@googlegroups.com
For more options, visit http://groups.google.com/group/mvdbms

[Will Johnson]

http://www.secnetix.de/olli/Python/block_indentation.hawk

[Peter McMurray]

Thanks. If only Dick had given a free education licence our world could be very different :-(

As for the indentation Hooray!. I have had to maintain some awful code with people writing 400 or 500 character lines -Yes I removed them pronto - or chopping about so much that I ran the program through B/list first.

 I also remember a chap who put  his IF  on a separate line from the actual command. Aaaargh!

I am delighted that the much maligned ED handles tabs properly - as per the Python explanation

"Most good editors support transparent translation of tabs, automatic indent and dedent. That is, when you press the tab key, the editor will insert enough spaces (not actual tab characters!) to get you to the next position which is a multiple of eight (or four, or whatever you prefer), and some other key (usually Backspace) will get you back to the previous indentation level."

apparently some Unix editors do not!

I might just have a closer look.

[Tony Gravagno]

If you're using ED to type Python... you're missing the point.

[Peter McMurray]

Hi TG It seems you are missing the point. 

My reference to the handling of tabs was purely to counter the poster who complained because some editors do not handle tabs properly

The point of my topic ii what does Python have to offer Pick users?

So far it seems to be Object orientation and simplicity of layout combined with the power of the Pick storage methodology PLUS a major training component.

If we could just get system programmers off their obsession with 16 bit - totally acceptable as a system programming methodology - and get them to realise that 8 bit data Unicode is the best data storage methodology, then Pick could boom again. Jbase is the only one so far, QM was a great disappointment in this area and Rocket seem to be concentrating on merging D3 with U2 to the considerable detriment of D3 with major errors in the latest releases of 10.2.1...

[Dan McGrath]

Standardized libraries for many common tasks so you don't have to reinvent everything. E.g, with UniBasic I had to write my own levenshtein distance to add suggested autocorrect to town/street names on green screen system. In Python I could have just done:

import Levenshtein

result = Levenshtein.ratio(..., ...)

Easier support for many integrations with other vendor solutions without having to roll my own low level interface:

• https://pypi.python.org/pypi/py-mailinator
  
• https://pypi.org/project/google-cloud-translate/
  

A language people have probably used elsewhere so not everyone needs to be retaught a new proprietary system. Sure, it's sort of easy to learn but has enough gotchas and oddities to slow onboarded - and working on proprietary languages also makes retention harder since the skill isn't as widely transferable.

Access to more modern tools for code development that I can use everywhere else, such as

• Linters: https://www.pylint.org/
• Code Coverage: https://coverage.readthedocs.io/en/coverage-4.4.1/
• Unit testing frameworks: https://docs.python.org/3.4/library/unittest.html

Many other reasons? But then again, I'm biased :)

--

You received this message because you are subscribed to
the "Pick and MultiValue Databases" group.
To post, email to: mvd...@googlegroups.com

To unsubscribe, email to: mvdbms+unsubscribe@googlegroups.com

[David Knight]

Hi Pete,

Not really OT; but I in respect your statement that Rocket is merging U2 & d3:

I specifically raised that when Rocket was here in Sydney a few month back, and the answer was an emphatic NO. The products are too diverged to ever be merged into one. I asked because I'm not a fan of U2; having used d3 forever. The roadmap for d3 looked pretty good to me.

That said, I agree lack of support for Unicode is a major factor. Certainly for my global domination plans!

[Peter McMurray]

Thanks Dan, that is the sort of info I was looking for. Levenshtein I had never heard of but looked up and can see that if you have a list of places it could be useful. However! if anyone ever uses it please make sure that you have legitimate lists or allow alternates - I get furious when I go to the bother of ploughing through several screens to order a specialist book or similar and on the last screen is a compulsory field containing only US states and dependencies.

Hi David, unfortunately another lights and music trip to hospital prevented me going to the meeting in Melbourne this year and D3 into U2 was definitely raised as positive last year. I am a long term D3 devotee and can remember precisely why I dumped Universe on NT circa 1997. I started conversion and discovered some clown had changed the INPUT @  by embedding some terminal highlight code that wrecked my screens. When I raised it as a fault, an even bigger ninny, who obviously knew nothing about foreground and background commands, told me it was a feature. I was working on Unidata on a DEC farm at the time and had a client running Universe on a SUN, neither of which had this issue. However! my target market was Windows. I have a simple rule, one easily discovered wreck is normally indicative of deeper more expensive wrecks to come.

Why then have we got such a failure in  D3 10.2.1 (printer issues); 10.2.2 failure to compile programs if subroutines are not physically placed in the program before they are used (this was fixed years back) Apparently somebody in system development has never heard of mainline code calling subroutines placed conveniently out of the way.

[Tony Gravagno]

On Friday, August 4, 2017 at 2:55:43 PM UTC-7, Peter McMurray wrote:

Hi TG It seems you are missing the point. 

My reference to the handling of tabs was purely to counter the poster who complained because some editors do not handle tabs properly

Uh, yeah, that's what I responded to... no point missed.

 

The point of my topic ii what does Python have to offer Pick users?

I agree with the responses on That topic, had nothing to add.
 

 and Rocket seem to be concentrating on merging D3 with U2 to the considerable detriment of D3 with major errors in the latest releases of 10.2.1...

Agreed on errors in 10.2.1, but 10.2.2 has been available for 4 months.

What makes you think they're merging D3 and U2? To my understanding, development teams there have been eager to understand the other platforms, and they are responding as eagerly to questions like "why can't we have X in platform Y like it is in Z?". But cross-pollenation of ideas doesn't equal "merge". We're all free to speculate but you're phrasing that as a statement so I'd be interested what substantiates such a statement.

T

[Tony Gravagno]

<< Pulled the trigger on a response before seeing David's note and Peter's response.
nvm....

[Will Johnson]

I don't see Python as something to "offer Pick programmers"

Rather it's something to offer Pick establishments that are treading water looking at the impending retirement of all those "Pick  programmers"

[Tony Gravagno]

I agree with Will - you don't hear that from me often. ;)
I believe Rocket is thinking exactly along those lines :
- Don't bother with people who aren't going to budge beyond BASIC. C'mon now, if they're not there after 20 years with all kinds of interfaces to/from these platforms, it's not going to happen.
- Don't rely on training to get mainstream people into Pick. Education has been a hit/miss sub-industry for decades and there just isn't enough desire in the field to allow that option to save the industry.

So the people who are here won't do it. New people aren't coming in to learn BASIC. The only vector left is to make the platform look more mainstream.

The next approach for this, and I'm hoping they're up to it, is Marketing to the mainstream! I'm not getting the impression that they're prepared for that. I'm not getting the warm fuzzies that they realize that "someone" needs to do that and it's not going to be the existing reseller channel. In their own interest of survival they need to do real Marketing and get new people to develop apps with Python and other languages over this platform.

My personal confidence in that is low. They've known that they have needed to do this for a couple decades and they haven't done it well. They didn't need Python - they could have done this with Java and .NET. Someone at the top there needs to realize that technology isn't the answer to their woes, it's pro-active efforts to do the real hard work, which is getting out there into the world, approaching new audiences, and trying to re-engineer what people are thinking about. All of these "build it and they will come" approaches have fallen because the vendors lacked the will-power to take that next step - and then they blame the platform and sell the assets off to the next buyer.

T

[Peter McMurray]

Hi TG

D3 10.2.2 is the version that has the major flaw concerning compile crashing if a subroutine is not physically placed in the code before it is used - this was fixed many years ago so somebody has mucked up code versions. The current advice from the suppliers is to install 9.4.2 as the last reliable version, although I have been using 10.2.0 in house for some time now.

The merging of D3 and U2 was announced in Melbourne 18 months ago but apparently dismissed as too difficult since then.

As for Python I definitely see the point regarding fresh blood now I know a little more about it.