LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (February 2005)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Mon, 14 Feb 2005 14:15:42 -0300
Reply-To:     Hector Maletta <hmaletta@fibertel.com.ar>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Hector Maletta <hmaletta@fibertel.com.ar>
Subject:      Re: Supporting Dale Glaser's comment on documentation
Comments: To: "Oliver, Richard" <richard@spss.com>
In-Reply-To:  <78A4B464DDC4E541B4ADD6A167830F1202664551@hqemail4.spss.com>
Content-Type: text/plain; charset="us-ascii"

Thanks, Richard, from this view from inside, which is quite illuminating. However, let me insist that this still doesn't explain the example about INDEX referred previously by Adrian Barnett. In that instance, entropy is evident. In other cases, even in the syntax reference, I often find that the difficult issues more in need of an explanation (though perhaps seldom used by a novice) are missing both from the explanatory text and the examples in the manual. I cannot say the same of the Case Studies since I have not looked them up on every aspect, but for the Syntax reference manuals this has been true for ages: cryptic explanation, and examples just covering the trivial, leaving you craving for something you desperately need to have clearly explained.

Criticism may hurt, but think it is a happy circumstance that there is still room for improvement in SPSS. Otherwise, what would the clever types at SPSS Inc be doing but reissueing the same perfect version for ever and ever? They would soon be fired, I suppose. I hope threads like this one help keeping all of you happily on the job for a long time to come.

Hector

> -----Original Message----- > From: Oliver, Richard [mailto:richard@spss.com] > Sent: Monday, February 14, 2005 1:04 PM > To: Hector Maletta; SPSSX-L@LISTSERV.UGA.EDU > Subject: RE: Supporting Dale Glaser's comment on documentation > > > Starting with SPSS 14, the syntax reference material will be > included in the context-sensitive help system, as well as > still available as a separate PDF document. > > As for the "diminishing amount of information per command at > each successive version," I hope this really isn't the case. > Although it's true that detailed statistical examples for > Base, Advanced, and Regression Models procedures are no > longer included in the manuals, examples are provided for > most procedures in the Case Studies, available from the Help > menu or by right-clicking any table produced by the > procedure. Right-mouse context help is also available for > many terms in tables. > > With the notable exception of removing detailed statistical > examples from certain manuals, we have made no > conscious/deliberate attempt to abridge any of our > documentation. In many instances, I suspect it's more an > issue of finding the information, since it's not all in one > place -- an issue we're still working on. And one step in > that direction is integrating the syntax reference material > with the help system. > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf Of Hector Maletta > Sent: Monday, February 14, 2005 9:47 AM > To: SPSSX-L@LISTSERV.UGA.EDU > Subject: Re: Supporting Dale Glaser's comment on documentation > > I have no qualms with the FILTER command, but love TEMPORARY > too. In fact, I loved much more another victim of unnatural > selection: PROCESS IF. In older versions of SPSS you could > fuse TEMPORARY and the subsequent SELECT IF commands into > one, by means of the PROCESS IF command. Why this had to be > split into two commands escapes me. Of course TEMPORARY can > be followed by ANY transformation command, not only SELECT > IF, and this makes sense of a generic TEMPORARY command, but > PROCESS IF could exist as well, couldn't it? > > By the way, I support the criticisms about the entropy > process undergone by manuals embedding a diminishing amount > of information per command at each successive version. If > those manuals are nowadays distributed as PDF files in the > installation CD, what is the point of abridging them? I could > understand the desire to keep the printed manuals affordable. > For the printed version readers might have the choice of an > abridged (or more > elementary) cheaper version or a more expensive and full > fledged one, but for the PDF format the cost difference is negligible. > > I also wonder why the materials in the Syntax reference > manuals (and the algorithms, come to think of it) are not > included in the context-sensitive Help system, instead of > coming separately. > > Hector > > > -----Original Message----- > > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf > > Of Peck, Jon > > Sent: Monday, February 14, 2005 12:18 PM > > To: SPSSX-L@LISTSERV.UGA.EDU > > Subject: Re: Supporting Dale Glaser's comment on documentation > > > > > > Well, your mileage may differ, but the filter route was > chosen for the > > gui because we believed that it fit the usual work pattern > much better > > than a TEMPORARY SELECT IF. It is pretty common to choose a subset > > and then run procedures against it several times, perhaps > iteratively > > refining a model or exploring several different aspects of the > > dataset. > > For this, having to reissue the TEMPORARY and SELECT IF every time > > seems to be a bad fit. And this logic really has nothing > particularly > > to do with the gui. It would be just as true for a syntax user. > > > > There are some other benefits of the filter approach in the gui: > > - the Data Editor shows clearly which cases are and are not > being used > > at the moment > > - a variable label that shows the filter condition is > attached to the > > variable > > - the Notes table includes that generated variable label, so the > > procedure output is well documented. > > > > Regards, > > Jon Peck > > > > -----Original Message----- > > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf > > Of Doug Fuller > > Sent: Monday, February 14, 2005 8:55 AM > > To: SPSSX-L@LISTSERV.UGA.EDU > > Subject: Re: [SPSSX-L] Supporting Dale Glaser's comment on > > documentation > > > > >Final pain-in-the-butt example. Using menus the select a subset of > > >cases for analysis creates a complex syntax creating a > > filter variable. > > >The menus have shielded from you the much easier TEMPORARY > > command to > > >make transformations apply to the one procedure only. > > > > This objection is only valid if you already knew syntax to > begin with > > and are bypassing the GUI anyway. A novice who is only > using the GUI > > is more prone to making the mistakes associated with forgetting the > > TEMP command in the first place, e.g. SELECT IF. You'll > note that in > > the GUI, the "Data->Select Cases..." subcommand defaults to > filtering > > rather than deleting. This is most certainly to require > the user to > > make the explicit choice to remove cases rather than filter. > > Certainly a more favorable default behavior, don't you agree? > > > > Indeed, using TEMP and SELECT IF in concert with a particular > > procedure is easier in syntax versus the GUI, but you > really only save > > one line of syntax in the end. > > > > TEMP. > > SELECT IF. > > PROC. > > > > versus > > > > COMPUTE $filter. > > FILTER BY $filter. > > PROC. > > FILTER OFF. (maybe) > > > > If disk/memory space are critical (i.e. your system can't handle the > > overhead of an additional filter variable), then I would posit that > > you already use syntax and the point is moot. > > > > In short, faulting the GUI for the user's lack of motivation to read > > the manual is like blaming your wallet for getting lost. > > > > Best, > > -- > > Doug Fuller > > dfuller@wayne.edu > > Research Assistant, Wayne State University Department of Psychiatry > > and Behavioral Neurosciences > > > >


Back to: Top of message | Previous page | Main SPSSX-L page