Date: Fri, 20 Jun 2008 08:35:31 -0400
Reply-To: Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject: Re: SAS's Future
In-Reply-To: <CA8F89971ADA9F47A6C915BA2397844207B428A8@MAILBE2.westat.com>
Content-Type: text/plain; charset="us-ascii"
I inadvertently posted this reply to the list. Apologies for deviating
so far from the topic of the original thread (whatever that was at the
time).
S
-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Sigurd Hermansen
Sent: Thursday, June 19, 2008 10:49 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: SAS's Future
Stephen:
A couple of people in my group are working actively with SharePoint.
Great for document and file exchange using virtual MS Windows copy
operations.... Other webparts such as calendars, discussion boards, and
issue tracking work well almost out of the box.
We don't have Excel access to or from our Web data zones. Also, the jury
is still out on the question of whether importing data via Excel imports
in SharePoint creates a real database. At least I don't see anything
with the virtues of a relational database.
We are looking now at whether or not we can work exclusively within
SharePoint. I don't find methods supporting database operations such as
joins and unions. As with XML, a new programming environment that
supports Cobol record structures doesn't do much for me.
Perhaps I will appreciate Sharepoint lists more as I learn more about
their possibilities. Lists can represent tuples. What about metadata and
key integrity constraints, though. How do you suppose that they will
work in SharePoint? S
-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Stephen McDaniel
Sent: Thursday, June 19, 2008 1:41 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: SAS's Future
Sig,
If you have the opportunity to use Excel with SharePoint, the latest
versions integrate well and easily allow data entry and modification
from the web. Here is a high-level overview of the process to address
this long-standing issue:
1) Make initial layout of columns and a few rows of data in Excel
2) Use the Excel spreadshseet to create a list in SharePoint (Import
Spreadsheet)
3) You will now have your data in a real database- SQL Server!
4) Customize columns, specify input types, set default values,apply data
rules via the SharePoint
5) You now have multi-user enabled data input/updating (per row),
records of user update/modifications, etc.
6) You can use TSQL to run additional validation, data quality, etc.
7) People can edit via thin web client or spreadsheet type interface on
the web
Really a leap forward and easy enough that most advanced Excel users can
be taught to create and maintain these in an hour or two!
Regards,
Stephen
On Wed, Jun 18, 2008 at 7:17 PM, Sigurd Hermansen <HERMANS1@westat.com>
wrote:
> Amidst all of the talk about the virtues of MS Excel and its audience
> of 40-50 million, doesn't anyone see a problem with sloppy garbage-in
> databases with mixed type attributes, cross-attribute values,
> strangely denormalized database designs, and crazy quilt worksheets.
> Programming using natural languages hasn't gone anywhere in twenty
> years. I suspect that attempts to automate conversions of haphazard
> collections of worksheet cells into databases will fail as well.
>
> We spend a lot of time restructuring and reworking data that Excel
> users have entered into dumb and smart Excel worksheets. Excel works
> much better as a destination than a source. Web data-entry systems
> work far better. A small investment in more intelligent data capture
> pays big dividends. Simply put, Excel simply gives beginning through
> expert users too many opportunities to subvert a database. I don't
> expect SAS to solve that problem any more than I would expect the Army
> Corps of Engineers to prevent floods.
>
> I do hope that the Iowa floods will abate soon and that the victims of
> the floods will recover quickly. S
>
> -----Original Message-----
> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
> Mary
> Sent: Tuesday, June 17, 2008 4:54 PM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: Re: SAS's Future
>
>
> Microsoft is making a run for it with SQL Server Analysis Services,
> but when I last viewed a demo of it over a year ago it looked like the
> statistics was fairly elementary. However, since SQL Server already
> has a language (Transact SQL stored procedures), that actually is a
> lot like SAS, it adding an analytics layer makes it a competitor to
> SAS in the data mining arena. But I don't think that the average user
> would be able to learn SQL Server.
>
> The big market is all the Excel users out there- there are tons of
> cost accountants and financial analysts who have reached the limits of
> Excel's statistical capability, and want a statistical package that
> would integrate well with Excel. Yes, you are right; if SAS could
> recognize that there are 40-50 million Excel users out there, and
> developed SAS to be more user friendly with Excel, it has a huge
> potential market.
>
> -Mary
>
>
> ----- Original Message -----
> From: Stephen McDaniel
> To: SAS-L@LISTSERV.UGA.EDU
> Sent: Tuesday, June 17, 2008 2:48 PM
> Subject: Re: SAS's Future
>
>
> Great observations Mary. R + Python + MySQL can come close to what
> you describe with moderate Excel integration at best...
>
> I also agree Microsoft has a ways to go on data mining.
> Unfortunately, they will capture the very large business analyst
> crowd when they do. SAS could cut this out by realizing that 3-4
> million users is great but much smaller
> than the potential 40-50 million person audience around the world who
> live,
> breathe, and use Excel in amazing ways. By lowering pricing moving
> forward
> (as they have somewhat done with BI Server) and heavily pushing a
> unified
> and elegant solution to business analysts that is based on sound
> statistical
> capabilities will be a big hit...
>
> Office 2007 support is coming, but 9.2 (client server) has been
> delayed for years... Maybe later this year?
>
> Also, EG 4.2 and the Add-In will adopt much more of the ribbon and a
> much simpler UI.
>
> Regards,
> Stephen
>
> On Tue, Jun 17, 2008 at 11:32 AM, Mary <mlhoward@avalon.net> wrote:
>
> > Actually, I think SAS has more basic things it should be doing; its
> > interface to Microsoft Excel isn't very good, in that it doesn't
> allow > specification of variable type and lets its import guess
> based on the first
> > 5 rows. Also it is not currently supporting Office 2007, which it
> needs to
> > be doing. The ability to import all spreadsheets within a workbook
> at once > would also be a great help.
> >
> > I also notice that there are many areas in the SAS display manager
> that do
> > not behave the way Microsoft products behave. One example would be
> "find
> > and replace" when viewing a SAS data set, and just the general
"look
> and
> > feel" of SAS compared to Microsoft products. Many of the
functions
> in the
> > display manager just feel very clunky. And then there is the idea
> of
> > following Microsoft's trend of ribbons rather than menus.... >
> > Perhaps SAS is trying to compete on too many platforms; and may
just
> want
> > to focus on Microsoft, Unix, and IBM mainframe platforms, and just
> not worry > about Linux; the vast majority of users now have been
> trained on Microsoft
> > products and operating systems.
> >
> > Also having SAS be "free" to those teaching it would help- the
> reason Java
> > got into computer science in Universities was that it was available
> free,
> > and then it had an edge coming out of colleges in that many
> students were
> > trained in it. The next generation of computer science students
> will be
> > trained in .NET with C#, so I think that they will be more
Microsoft
> > oriented. Actually, the interest in JAVA on college campuses seems
> to be > dying out.
> >
> > SAS still has tremendous capabilities to merge and manipulate data,
> and
> > although I don't know R, I would anticipate that it has nothing
like
> the
> > capability yet of Proc SQL. That SAS can do this, and then use
> that data
> > in its statistical procedures is its strong suit. Also, SAS seems
> not to > understand that it is the ability to write programs in SAS
> that makes it
> > different from interactive packages; and ODS makes the ability to
> capture
> > pieces of output to save in data sets for later use or to summarize
> results
> > that makes it very powerful.
> >
> > Thus if SAS took the steps necessary to improve its display manager
> to
> > behave more like Microsoft Office, if it gave it away free to
> college
> > students, and developed its editor in ways that Excel has, such as
> fill-in
> > functions, I think it could have a long future.
> >
> > Excel is a long way from becoming a data mining tool; in that it
> still
> > cannot merge at all and its statistical functions are crude and
> elementary,
> > plus its VB Excel programming language leaves much to be desired.
> Thus, I
> > don't think it competes with SAS Base and SAS Stat, though
certainly
> does in
> > terms of graphics.
> >
> > If SAS could be more compatible with Excel, and have a similar
> interface to > Excel, then I think it can withstand the tide.
> >
> > -Mary
> >
> >
> > ----- Original Message -----
> > From: Alan Churchill
> > To: SAS-L@LISTSERV.UGA.EDU
> > Sent: Tuesday, June 17, 2008 11:11 AM
> > Subject: Re: OT Friday Humor
> >
> >
> > SCL requires Base, does it not?
> >
> > .NET is free, is open source, and runs on Linux as well as
Windows.
> > Silverlight, a .NET subset, also runs on Macs. You also have Java
> and perl > as choices (as well as loads of others).
> >
> > How do you justify SCL when it requires thousands of dollars per
> desktop
> > for
> > deployment? I don't have an issue with SCL but SAS needs to invest
> money
> > into it and make deployment a no cost issue. I doubt that is their
> plan
> > though since they are investing heavily in Java and .NET.
> >
> > Alan
> >
> > Alan Churchill
> > Savian
> > www.savian.net
> >
> >
> > -----Original Message-----
> > From: Dave Scocca [mailto:dave@scocca.org]
> > Sent: Tuesday, June 17, 2008 8:22 AM
> > To: Alan Churchill; SAS(r) Discussion
> > Subject: Re: [SAS-L] OT Friday Humor
> >
> >
> > --On 6/17/2008 8:19 AM -0600 Alan Churchill wrote:
> >
> > > [...].NET [...] all have 0 cost
> > > deployments.
> >
> > Yeah, because Microsoft gives Windows away for free.
> >
> > D.
> >
>
| 
