UNIX vs NT read the facts before deciding which way to go
Todd
Diego there were no plans to support Linux...perhaps he should reconsider.
Go to URL and see why UNIX in general and Linux in particular is better than
NT.
tur...@my-dejanews.com
"Todd" <ne...@webshaker.com> wrote:
> According to Jack Dangermond of ESRI at the 1998 Users Cconference in San
> Diego there were no plans to support Linux...perhaps he should reconsider.
The two products that would make a lot of sense are SDE and ARC/INFO. Now
that the major RDBMS manufacturers have ported their databases to Linux, SDE
would be a match made in heaven. At work we have SDE running on an Alpha
connected to Oracle on Linux. It works well. Surely something as simple as
SDE would be no more than a couple of days recompile and testing?
ARC/INFO would be great too. It'd be the best of both worlds - low cost
hardware and the familiar, wonderful UNIX environment. A/I just doesn't work
well in NT - I hate it. Again, we're talking UNIX to UNIX here - how much
effort would it be to port A/I?
Of course, it'd be a major undertaking to _support_ Linux and maybe doesn't
have any Linux experts. But then, A/I and SDE already run on more obscure
UNIXes such as HP/UX and Digital Unix.
I'd doubt if any client-side technology like ArcView or MapObjects would ever
be ported to Linux. ESRI has already made it clear they are pulling out of
the UNIX space for ArcView and inevitably, I think, A/I and SDE will follow.
Jack has made his pact with the devil. Sad, really.
Here's hoping that WINE (http://www.winehq.com) will allow me to, one day, run
ArcView on my Linux box.
> Go to URL and see why UNIX in general and Linux in particular is better than
> NT.
yeah, yeah. We all know that already but it doesn't seem to make a
difference.
-t.
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Marti Rijken
>
>Go to URL and see why UNIX in general and Linux in particular is better than
>NT.
Better or not is not the issue, I am afraid. It is the market potential.
But then, the amazingly fast growing number of Linux users should be
reason enough to reconsider.
A personal note added: ArcView on W'95 sucks so badly, that we have been
forced to increase our number of ArcView-HPUX licenses for serious tasks
(i.e. more than just viewing data).
Cheers,
--
@..@ Marti Rijken <mri...@natrix.demon.nl>
(`--') at work: mri...@prv.gelderland.nl
( }..{ ) URL: http://www.natrix.demon.nl/
Allan Doyle
available Java toolkit for which you also get sources. OpenMap can deal
with Shapefiles (but we don't support arcs yet). If anyone wants to help
finish the Shapefile reading, we'd be happy to hear from you. If there's
no perceived market for ESRI in Linux, maybe ESRI might even want to
donate some Java code to this effort?
Note that OpenMap is a toolkit, not an end-user tool. On the other hand,
that's why we built it in the first place. We use it to build a variety
of end-user tools...
Allan Doyle
http://openmap.bbn.com
Todd wrote:
>
> According to Jack Dangermond of ESRI at the 1998 Users Cconference in San
> Diego there were no plans to support Linux...perhaps he should reconsider.
>
> Go to URL and see why UNIX in general and Linux in particular is better than
> NT.
>
> http://www.unix-vs-nt.org
>
> Todd
> www.webshaker.com
--
______________________________________________________
see http://openmap.bbn.com for OpenMap(tm) info.
Use openmap@spam_me_not.bbn.com for OpenMap questions
remove the "spam_me_not." part first!
OpenMap is a trademark of BBN Corporation
Todd
>Jack has made his pact with the devil. Sad, really.
Hmm Bill Gates is Satan? Never would have guessed =o)
Dimitri Rotow
wrote:
>On Mon, 4 Jan 1999 18:18:45 -0400, Todd <ne...@webshaker.com> wrote:
>>
>>Go to URL and see why UNIX in general and Linux in particular is better than
>>NT.
>
>Better or not is not the issue, I am afraid. It is the market potential.
>But then, the amazingly fast growing number of Linux users should be
>reason enough to reconsider.
>
Let me add a pro-Windows perspective. I've been working with UNIX
since the mid 1970's and was the general manager at Intel responsible
for making available a very low cost UNIX port on Intel processors.
Getting UNIX onto Intel processors certainly played a key role in
getting UNIX volume numbers up, so one cannot accuse me of failing to
pull my oar in moving UNIX forward.
One would think a sub-$100 UNIX and sub-$1000 hardware on which to run
it would be a cool idea. It is, but it did not take off the way
Windows did, nor is it taking off now. Linux is growing fast only
as a result of the growth of embedded appliance usage within servers.
That's not where applications standards are won. In mass markets,
Windows is growing much faster.
I gave up on UNIX mainly because UNIX was falling behind and becoming
increasingly difficult to develop for or to use in real life. I think
this occured because UNIX was owned and steered by either a) a
brain-dead computer company (AT&T) or b) a committee. One simply
could not count on either "owner" to deliver realistic and effective
approaches to new computing markets characterized by PCs that were as
good as Microsoft's offerings. I admit it. As a user I like the idea
of cheap software that works great with a choice of millions of
applications, many tens of thousands of which are free. The gateway
to all that is parting with a few dollars for Mr. Bill. No problem!
Windows is probably the best deal in computing today [well, after
Manifold...:-) ]
Neither a brain-dead company nor a design by committee is a good bet
when competing with a team of fanatic, technically-skilled people who
really understand PC markets. Bill Gates happens to be one tough cat
who also personally understands the technology and business of PC
software and hardware better than nearly anyone else in the world.
Steve Ballmer is not only smart and effective, he is also a good guy.
While the legendary turkeys at AT&T were missing one opportunity after
another, the team at Microsoft was steadily improving the value of
their product to consumers. Give them credit for how far they have
come.
Note that while SUN also is run by pretty smart guys, they have not
owned UNIX nor have they ever been able to bust out of a 70's style
"proprietary minicomputer" approach to the market. This is an uphill
sell against the astonishly open PC clone standard, and marketing talk
that attempts to position proprietary deals as "open" standards is not
enough (as SUN learned with SPARC). Tying UNIX to SPARC as SUN did is
more than enough to sink it in the mass market no matter how well it
worked in SUN's niche markets.
People in UNIX have become fond of complaining about Microsoft's
stick, but I've heard no recognition of the carrot in the form of
spectacular development tools. Bill was much smarter than his UNIX
business counterparts because he realized years ago that the future of
an OS was tied to applications, and applications are tied most of all
to the development tools. So, he invested big time in creating
elegant, easy to use, dirt cheap development tools for standard
languages that delivered everything a developer could ever want to
enter the mass market.
It's not just that developing for Microsoft means you automatically
get a market that's 100 times larger than UNIX... you also get to use
development tools that are far cheaper, more efficient, and deliver
superior integration with operating system and applications. This is a
surprising reversal of roles. Ten years ago the UNIX dev tools were
better, albeit much more expensive. Now, the Windows tools are much
better.
Look, I think Linux is great, but there are no development tools
within Linux comparable as a total solution to Visual Studio, VC, MFC
and all the other fun Microsoft stuff. If we develop in Visual, we
get automatic access to, well, Access, and all the other tens of
thousands of applications created with Jet. In a modern application
like Manifold, one needs not only database connectivity to a modern
(ahem... that term does *not* include dbf) native database, but to
ODBC as well. One needs automatic interface to networking protocols,
integrated help system, professional source code control, language
support for scripting language standards such as VB and Javascript,
execution of Internet browsing, image viewing, sound recording, and
much, much more.
Personally, I think a modern GIS needs to do much more than simply
draw bits in a window. Manifold, for example, will let you point and
click at spot on your map and "tag" it with a voice note. Someone
later can click on it and hear your spoken commentary. Within
Windows, this is almost free for us to do [plus, automatic support for
a few thousand different sound cards]. The capability to do this is
almost automatic within Visual. Within Linux? well.... it's doable,
but not so easy.
I suppose it's possible to piece together an analogous development
environment in UNIX, but at what cost? Who's going to support all the
different bits and pieces needed to simply approach what one gets in a
low-cost, efficient package from Microsoft? Where is the VBA
connectivity to the applications used by 99% of the planet? I
suppose there are a few people out there who are still waiting for
Avenue or Forth to sweep the planet as a modern language standard, but
most people these days script applications in VB and reasonably expect
to do the same in their GIS package.
One reads much criticism of various legacy GIS packages in this
newsgroup. The common thread, it seems to me, is they have been
developed in an earlier time and are suffering from ancient approaches
to computing. Half the time I read the complaints about a legacy GIS
and it sounds like people are complaining about issues related to
UNIX, such as the persistance of command line interfaces.
UNIX simply has failed to evolve the infinity of cool user interface
stuff we take for granted in Windows environments. If you develop for
the latest Windows environments using the latest Windows tools, you
can deliver to your GIS customer a spectacular array of capability at
a very low price. One can attempt to do the same in UNIX but the
result is less capability, slower evolution, and a price ten times
higher.
Let's face it: the UNIX of today is succeeding based mainly on the
technical accomplishments of an earlier day that produced a first-rate
multi-user kernel and a set of legacy networking protocols. That
explains the popularity of Linux and Free BSD for running Web servers.
But on the other side, no one has ever accused UNIX of delivering a
highly-evolved, state of the art user interface that might appeal to
the masses. Attempts to evolve UNIX in this direction have lagged
far behind progress in the PC community. Sure, there are PC
applications that have been ported to UNIX, but so far one sees
second-rate versions appearing in UNIX that are quite limited compared
to the latest and greatest stuff done within Windows.
If the UNIX community wants to move ahead of Microsoft in mass
markets, it's time to forget the politics, stop calling Bill names and
to get on with a program of *exceeding* the value, quality, and range
of the development tools fielded by Microsoft, especially as regards
user interfaces. Unless this happens, UNIX will be limited to three
markets: hobby use, proprietary workstations, and a very steady
"embedded" market within anonymous server appliances.
Cheers,
Dimitri
PS: hey, this isn't one of those newsgroups where a pro-Windows rant
draws flames from UNIX partisans, is it? :-)
Manifold Net Ltd
1945 North Carson St. Suite 700 Phone: 800-556-5919
Carson City, Nevada 89701 USA
Bob Taylor
Dimitri Rotow wrote:
> On 5 Jan 1999 07:32:47 GMT, mri...@natrix.gld.nl (Marti Rijken)
> wrote:
>
> >On Mon, 4 Jan 1999 18:18:45 -0400, Todd <ne...@webshaker.com> wrote:
> >>
> >>Go to URL and see why UNIX in general and Linux in particular is better than
> >>NT.
> >
> >Better or not is not the issue, I am afraid. It is the market potential.
> >But then, the amazingly fast growing number of Linux users should be
> >reason enough to reconsider.
> >
>
> Let me add a pro-Windows perspective. I've been working with UNIX
> since the mid 1970's and was the general manager at Intel responsible
> for making available a very low cost UNIX port on Intel processors.
> Getting UNIX onto Intel processors certainly played a key role in
> getting UNIX volume numbers up, so one cannot accuse me of failing to
> pull my oar in moving UNIX forward.
>
>
<snip?
> If the UNIX community wants to move ahead of Microsoft in mass
> markets, it's time to forget the politics, stop calling Bill names and
> to get on with a program of *exceeding* the value, quality, and range
> of the development tools fielded by Microsoft, especially as regards
> user interfaces. Unless this happens, UNIX will be limited to three
> markets: hobby use, proprietary workstations, and a very steady
> "embedded" market within anonymous server appliances.
>
> Cheers,
>
> Dimitri
>
Fair enough, for those whose primary focus is the interface. But others of us
analyze big databases and sometimes tire of doing something three times before it
works.
Bob
Dimitri Rotow
<beaucoup snip>
>Fair enough, for those whose primary focus is the interface. But others of us
>analyze big databases and sometimes tire of doing something three times before it
>works.
>
>Bob
>
Depends on what you mean by "big". If you are talking terabytes, you
certainly have a point. But this is a pretty rarified niche.
If we are talking the usual run of dozens of gigabytes and such, NT
would be a fine choice, as would UNIX. I'm not sure why you would
have to repeat things. Pairing NT 4.0 with databases such as Oracle
and current SQL Server releases, there's no way you'd have to do
something three times. We run SQL Server with 4.0 and have no
problems, so I don't understand the comment. Perhaps it is a database
issue, version skew during an upgrade, or some other problem.
By the way, for database work there are many benefits for developers
with Microsoft. We are doing a lot of work in database these days,
and the appeal of interfacing to SQL Server, Plato (Microsoft's OLAP)
and similar development support makes this a very attractive
environment for sophisticated analysis of databases into the multi-
gigabyte range.
Cheers,
Dimitri
Mark Cederholm
>with Microsoft. We are doing a lot of work in database these days,
>and the appeal of interfacing to SQL Server, Plato (Microsoft's OLAP)
I'm not so fond of SQL per se (like Java, it's not so much a standard as a
point of departure), but I'm starting to play with DAO in Visual C++ right
now and it's fun! I think Unix could use something like that.
-- Mark
************************************************************
"We must cease confusing mastery of software commands with
attaining a grasp of critical intellectual concepts."
-- Duane F. Marble
************************************************************
Dan
Dimitri Rotow wrote in message <36925918...@news.dnai.com>... <big snip>
One reads much criticism of various legacy GIS packages in this
newsgroup. The common thread, it seems to me, is they have been
developed in an earlier time and are suffering from ancient approaches
to computing. Half the time I read the complaints about a legacy GIS
and it sounds like people are complaining about issues related to
UNIX, such as the persistance of command line interfaces.
<another big snip>
You make a very good point here. I think its a knee jerk reaction to
blame the OS, as was done in this thread with ArcInfo for NT. Is it NT or
is it the port of an application that was developed on a different OS (i.e.
UNIX)? Over time these bugs will be worked out I'm sure, but you just can't
take a square and expect it to fit in a circle without the proper
modifications. I watched the Intergraph GIS products get ported to NT from
CLIX (yet another form of UNIX) with a great deal of pain and effort. But
with time and effort persistence paid off!
Dan
Eric Miller
> Look, I think Linux is great, but there are no development tools
> within Linux comparable as a total solution to Visual Studio, VC, MFC
> and all the other fun Microsoft stuff. If we develop in Visual, we
> get automatic access to, well, Access, and all the other tens of
> thousands of applications created with Jet. In a modern application
> like Manifold, one needs not only database connectivity to a modern
> (ahem... that term does *not* include dbf) native database, but to
> ODBC as well. One needs automatic interface to networking protocols,
> integrated help system, professional source code control, language
> support for scripting language standards such as VB and Javascript,
> execution of Internet browsing, image viewing, sound recording, and
> much, much more.
I'd have to agree with you that it's a bit more difficult to develop in
Linux/UNIX. There are several development tools and toolkits however.
The gnome project, www.gnome.org, which will have its 1.0 release soon,
and GTK, www.gtk.org, will help to bridge some of the consistency and
interoperability issues. (GNOME uses a CORBA implementation called ORBit
to provide an interface for services). Emacs has many modes for writing
in different languages, with things like syntax highlighting, though the
WYSIWYG interface environments aren't there yet. If you know Perl, which
is much more portable than VB, you can write simple tight interfaces to
a variety of databases or other applications using application specific
bindings others have written or more general ones like DBD.pm/DBI.pm.
There are also quite a few java tools/toolkits/IDLs becoming available
that run on linux.
It's obvious though, that in it's current state, Linux has more to offer
in server or embedded systems scenarios than as a general purpose
workstation. But the price and performance for those scenarios is hard
to beat.
--
Eric Miller: eg...@tidepool.com http://www.tidepool.com/~egm2
There once was a fellow named Pope,
Who plugged into an oscilloscope.
The cyclical trace
Of their carnal embrace
Had a damn near infinite slope.
-- Thomas Pynchon
Olga A. Blinkova
understand why. There are three distinguishing features
of Manifold: 1) integration with Access (i.e. MS Office);
2) ability to develop VB applications; 3) cool mathematics
and statistics. So I highly doubt MS MapPoint will receive 80%
of GIS market, as it was said. The price is almost the same
($109 vs $107), Manifold is much more powerful product than
MapPoint, Manifold is integrated into Office. It looks like
Manifold is a new leader of GIS market...
Olga Blinkova,
"GIS-Review" (Moscow)