Date: Tue, 16 Mar 2010 10:33:53 -0600
Reply-To: Alan Churchill <alan.churchill@SAVIAN.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Alan Churchill <alan.churchill@SAVIAN.NET>
Subject: Re: DDE and macro recall
In-Reply-To: <20100316090311.542C5C6B@resin16.mta.everyone.net>
Content-Type: text/plain; charset="utf-8"
Yes but that wasn't what brought SAS into prominence. I think it was the language ease that mattered.
I remember, back in the day, when I went to a SAS class in Austin on high data volume and how to use SAS. The instructor was talking about 100k obs. My buddy and I chuckled because we were working at MCI at the time and were analyzing all calls into MCI customer service. Well, that was a massive volume. 100k records was a fraction of our test cases. We got nothing out of the class because we were already using their most advanced techniques and had actually gone beyond them.
At that time, we did SyncSort --> Easytrieve --> SAS to cope with the number of records. We didn't use SAS up front due to efficiency (SyncSort smoked it 6 ways to Sunday). We did use it on the final push due to ease of the language for ETL work.
Now we have the new kids on the block (Java, C#, R, perl, python, etc.) who all want to capture SAS programmers' hearts. SAS programmers, though, are different. From everything I have seen over these decades, a lot of SAS coders seem to have started with the business side first and then went to SAS because they needed more flexibility. Hard-core coders pop in, no doubt, but it doesn't seem like the SAS masses are of that ilk.
Record volume isn't that big of a deal anymore, IMO. I can actually use C# paired with SQL Server, Oracle, DB2 and scream through data volumes. There is something else there and I think it is simple ease of use for a business person.
My 1 cent and I am sticking with it ;-]
Alan
Alan Churchill
Savian
www.savian.net
Office: (719) 687-5954
Cell: (719) 310-4870
-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Mary
Sent: Tuesday, March 16, 2010 10:03 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: DDE and macro recall
I think the reason that SAS users don't use a low level language is that the nature of what they are doing is often dealing with large amounts of data to do aggregation, reporting, or statistical analysis. R simply can't handle millions of records and this is why such tasks are done in SAS.
I did previously work for a shop that had PL/1 and SAS, and did program in PL/1 principally because access to DB2 was restricted via licensing to PL/1 and was not available on SAS at that particular shop, but many shops do have access to the database setup so that one can access the database directly though SAS's Proc SQL. There wasn't really much in terms of code in PL/1 that I couldn't accomplish in SAS, so there was no main other reason to program in PL/1.
On the other hand, applications that have GUI interfaces are unsuited to SAS. I programmed in PowerBuilder and building that sort of front-end requires a language that is designed for it- with objects for drop-down boxes, folders, command buttons, and the like. This is a totally different sort of programming than most SAS programmers do, and thus requires different tools.
-Mary
--- alan.churchill@SAVIAN.NET wrote:
From: Alan Churchill <alan.churchill@SAVIAN.NET>
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: DDE and macro recall
Date: Tue, 16 Mar 2010 08:38:25 -0600
Art,
I think there are a number of SAS users who would entertain using low-level
languages but not most. I have simply been to too many sites and have seen a
lot of SAS users and they shy away from non-SAS languages. We even see that
on the L. This isn't a disparagement. SAS users tend to need business info
and need a simple way to get to it to make business decisions. Diving deep
into comp sci isn't their thing.
Can SAS users do this? Sure. I just don't believe that most want to or feel
a need to.
Hope that clarifies my position.
Alan
Alan Churchill
Savian
www.savian.net
Office: (719) 687-5954
Cell: (719) 310-4870
-----Original Message-----
From: Arthur Tabachneck [mailto:art297@NETSCAPE.NET]
Sent: Tuesday, March 16, 2010 8:27 AM
To: SAS-L@LISTSERV.UGA.EDU; Alan Churchill
Subject: Re: DDE and macro recall
Alan,
With respect to "I do not think most SAS users would use perl.", I
distinctly remember at least one quite prominent one. See:
http://www.listserv.uga.edu/cgi-bin/wa?S2=sas-
l&D=1&H=0&O=D&T=1&q=perl&s=&f=cassell&a=&b=
or, in short form, http://xrl.us/bgys3f
Art
-------
On Tue, 16 Mar 2010 01:32:24 -0600, Alan Churchill
<alan.churchill@SAVIAN.NET> wrote:
>HoboCopy looks like an interesting utility. This may come in handy for
>COM locked apps like Office.
>
>Insofar as R and perl are concerned, I do not think most SAS users
>would
use
>perl. It is simply way too low-level. I find similar resistance to C#
which
>is probably more friendly due to Visual Studios and pure OOP.
>
>R is only needed by the stat guys and they tend to like what they are
>accustomed to (from my experience).
>
>I encourage alternative technologies because it helps on lots of fronts
but
>the above is simply my observations after working with SAS users for
>20+ years.
>
>Alan
>
>Alan Churchill
>Savian
>www.savian.net
>Office: (719) 687-5954
>Cell: (719) 310-4870
>
>
>-----Original Message-----
>From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
>xlr82sas
>Sent: Monday, March 15, 2010 2:06 PM
>To: SAS-L@LISTSERV.UGA.EDU
>Subject: Re: DDE and macro recall
>
>On Mar 15, 9:54 am, montura <montura...@gmail.com> wrote:
>> PERL and R?
>> You must drunk. Who the hell needs that?
>
>Hi,
>
>I am not sure where Microsoft is going with DDE(see below), but this
>may interest some readers.
>
>What happened to DDE Share?
>
>Applies to all editions of Windows Vista
>
>If you used previous versions of Windows, you might know DDE Share as a
tool
>for managing the way programs communicate and share data over a network.
DDE
>Share is not available in this version of Windows because it has been
>replaced by other methods for communicating and sharing data among
>computers.
>
>My concern:
>=========
>In the past there was no way to copy a networked shared excel
>spreadsheet while multiple users were updating the file, this is why I
needed DDE.
>Vista has something called hobocopy? to copy 'exclusive open' files? I
>am investigating.
>
>
==========================================================================
>===========================
>Also the list of R users now has some very heavey weight 'past' SAS
>power users? Note Micheal Freindly, not to mention Frank Harrel and others.
>They can't all be wrong.
>
>The list below is only about 1/5th of the 'new/updated' packages in R
>last week.
>
>UScensus2000blkgrp (0.03)
> Zack W. Almquist
> http://crantastic.org/packages/UScensus2000blkgrp
>
> US Census 2000 Block Group shapefiles and additional demographic data
> from the SF1 100 percent files. This data set contains polygon files
> in lat/lon coordinates and the corresponding demographic data for a
> number of different variables.
>
>
> Michael Friendly
> http://crantastic.org/packages/vcdExtra
>
> Provides additional data sets, methods and documentation to
> complement the vcd package for Visualizing Categorical Data.
>
>Rsolnp (1.0-2)
> Alexios Ghalanos
> http://crantastic.org/packages/Rsolnp
>
> General Non-linear Optimization Using Augmented Lagrange Multiplier
> Method
>
>* saws (0.9-3.1)
> M.P. Fay
> http://crantastic.org/packages/saws
>
> Tests coefficients with sandwich estimator of variance and with small
> samples. Regression types supported are gee, cox regression, and
> conditional logistic regression.
>
>RcppExamples (0.1.0)
> Dirk Eddelbuettel and Romain Francois
> http://crantastic.org/packages/RcppExamples
>
> Examples for Seamless R and C++ integration The Rcpp package contains
> a C++ library that facilitates the integration of R and C++ in
> various ways. This package provides examples.
>
>RcmdrPlugin.MAd (0.2)
> AC Del Re
> http://crantastic.org/packages/RcmdrPlugin-MAd
>
> This is an R-Commander plug-in for the MAd package (Meta-Analysis
> with Mean Differences). This package enables the user to conduct a
> meta-analysis in a menu-driven, graphical user interface environment
> (e.g., SPSS), while having the full statistical capabilities of R and
> the MAd package. The MAd package itself contains a variety of useful
> functions for conducting a research synthesis with mean differences
> data. One of the unique features of the MAd package is in its
> integration of user-friendly functions to complete many of the
> statistical steps involved in a meta-analysis with mean differences.
> It uses recommended procedures as described in The Handbook of
> Research Synthesis and Meta-Analysis (Cooper, Hedges, & Valentine,
> 2009).
>
>
> ordinal (2010.03-04)
> Rune Haubo B Christensen
> http://crantastic.org/packages/ordinal
>
> This package implements likelihood based models for ordinal (ordered
> categorical) data based on cumulative probabilities in the framework
> of cumulative link models. This includes the important proportional
> odds model but also allows for general regression structures for
> location as well as scale of the latent distribution, i.e. additive
> as well as multiplicative structures, structured thresholds
> (cut-points), nominal effects and flexible link functions. Further, a
> range of estimation procedures and a range of auxiliary functions are
> implemented.
>
>mtsc (0.0.1)
> Charlotte Maia
> http://crantastic.org/packages/mtsc
>
> Place-holder package (roughly speaking an empty package) for finding
> clusters in multivariate timeseries. Full implementation pending.
>
>MplusAutomation (0.2-3)
> Michael Hallquist
> http://crantastic.org/packages/MplusAutomation
>
> The MplusAutomation package leverages the flexibility of the R
> language to automate latent variable model estimation and
> interpretation using Mplus, a powerful latent variable modeling
> program developed by Muthen and Muthen (www.statmodel.com).
> Specifically, MplusAutomation provides routines for creating related
> groups of models, running batches of models, and extracting and
> tabulating model parameters and fit statistics.
