Re: It is a problem that Windows PSPPIRE/PSPP is pretty wonky

Alan,

I understand your point now. Sorry for misunderstanding.
I agree with you but same as you I don't see any solution yet. If there were a active PSPP developer who uses Windows this situation would probably look different. I guess it would be hard to find such new developer to join the team, but maybe the Windows using tester or testers would be enough? I'm not a tester but I'm learning programming in Python so I understand some and start to understand more and more computer science concepts and I could learn more. I have access to Windows 7 (32 and 64 bit) and could compare results with PSPP version running under Linux. I guess we could find some more people ready to perform tests under Windows if you think that it could help.

kind regards,

Michal

2014-12-31 18:05 GMT+01:00 Alan Mead <address@hidden>:

Michal,

You are preaching to the choir; one of the ways PSPP improves is through reports about the wonky bits.

However I had a point which I didn't make explicitly that bears explication: I'm not at all sure that PSPP FOR WINDOWS is under active development. One of the ways (probably the main way) that free software gets better is when developers are motivated to solve their own problems. Recently Ben added support for new encrypted SAV files and even more recently old SPSS+ SAV formats. He did that because he enjoys that kind of work. That's how free software works and why it's often better than commercial software.

And here's the point: I don't think that Ben or John ever use Windows, so any annoyances specific to the Windows version (like not being able to paste properly) are invisible to them and will never be fixed. Windows is treated like a distribution package and the Windows-specific issues are a problem for a Windows package manager to solve. But there is a gap there. I am grateful for (or I admire and applaud) Harry's work to provide a Windows version, but he cross-compiles it on (I think) OpenSUSE. I don't know that there is any developer of PSPP who regularly uses PSPP on Windows and who is motivated to solve these annoyances in the Windows version.

And I think that's a problem because, despite Sinofsky's best efforts to drive everyone away, Windows still has a desktop market share around ~90% and I'd expect that means ~90% of new PSPP users are trying to use PSPP's second-class implementation, the Windows version. My specific motivation for posting about the wonky bits was as a bug report; a Windows user was trying to use PSPP on Windows to make boxplots and I had pasted exact steps but even if that user uses the right steps they will encounter the paste problem and the boxplot they get will (presumably) be flawed in the ways I described. A couple weeks ago, I had a new PSPP user email me off list because she was running Windows 8 and she reported that PSPP won't run at all on her computer. I wouldn't be surprised to find out that she was the first to try PSPP on Windows 8 (I know I don't have a Windows 8 machine to try). Another problem I've encountered is that it's easy to create multiple instances of PSPP(IRE) on Windows (just double-click on two or more different SAV files in Windows Explorer). So, you can easily have a situation where you have a dataset open in one window and syntax in another but, because the windows are attached to independent instances of PSPP, the syntax cannot access that data. To solve this, one must open a new syntax window from the window containing target the dataset and paste your syntax into that window (or save the syntax from one window to a file and open it in the other window). For an SPSS user, this is extremely confusing because in SPSS any syntax window can activate any dataset; it's probably fairly confusing to people who've never used SPSS. But the overall issue I'm addressing is that I think it's a problem for PSPP that the Windows version is so wonky.

So, as you say, this is free software and there is no solution per se, but I think this is an appropriate forum for raising this issue.

-Alan

On 12/31/2014 9:09 AM, Michał Dubrawski wrote:

Dear Alan,

Although what you are saying are things that could be improved in PSPP you should remember that PSPP is still under development - it is not version 1.0.
-- 

Alan D. Mead, Ph.D.
President, Talent Algorithms Inc.

science + technology = better workers

+815.588.3846 (Office)
+267.334.4143 (Mobile)

http://www.alanmead.org

Announcing the Journal of Computerized Adaptive Testing (JCAT), a
peer-reviewed electronic journal designed to advance the science and
practice of computerized adaptive testing: http://www.iacat.org/jcat

From:	Alan Mead
Subject:	Re: It is a problem that Windows PSPPIRE/PSPP is pretty wonky
Date:	Wed, 31 Dec 2014 11:58:12 -0600
User-agent:	Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0