GISUser.com has a news release posted that probably was meant to go out tomorrow, but “today” ArcGIS 9.3 ships. (HT Bill Dollins)
So now go bug your local ESRI rep asking them when it will show up.
Caption not needed
ESRI ArcGIS 9.3 Ships Tomorrow
June 25th, 2008 · 75 Comments · GIS
Tags: arceditor·arcgis·arcgis 9.3·ArcGIS Desktop·ArcGIS Engine·ArcGIS Server·ArcIMS·arcinfo·ArcObjects·ArcSDE·arcview·ESRI
75 responses so far ↓
1
Jason Stewart
// Jun 25, 2008 at 9:26 pm
Just checked on
http://esri.com/software/arcgis/index.html
and guess what it says:
“ArcGIS 9.3 is now available. Learn more.”
Cheers.
It’s 2:30 pm on the 26th of June, here in Australia, anyway!
2
Friend of GIS
// Jun 26, 2008 at 5:32 am
Why do you think the release was posted a day early?
3
Joseph Wallis
// Jun 26, 2008 at 7:13 am
actually it was already shipping as of last week.
4
Drew
// Jun 26, 2008 at 7:32 am
As for the all important web help - nowhere to be seen as of yet. Still in beta lock down mode. No idea why though. Any thoughts?
5
Simon
// Jun 26, 2008 at 7:42 am
esri uk informs me that im on the priority list but to expect it in 2-3 weeks
6
Gady
// Jun 26, 2008 at 7:53 am
Too bad the Resource Center is still asking for beta passcode. I want my JavaScript API!
7
Chris M
// Jun 26, 2008 at 9:19 am
I got the RC version a couple of weeks ago but didn’t get a chance to play with due to vacation and workload. However, I did her from some developers that it seemed kind of buggy. I hope this is not the case as I am getting ready to create a new GDB on a SQL server and would use 9.3 from the start then install 9.2 and have to go through the migration progress later.
8
JHall
// Jun 26, 2008 at 1:35 pm
Chris, the user and developer community has no one but ourselve to blame, if we don’t take the time to test the RC and report to ESRI any problems found.
We did “extensive” (several days) testing of the release candidate, after having problems in the past with ESRI’s major releases or service packs breaking ArcObjects interfaces that previously worked fine in our ArcGIS Desktop extension.
The good news was that we found no problems in the RC . I am cautiously optimistic, this time around.
9
jxn
// Jun 26, 2008 at 4:49 pm
What are your thoughts on when is the right time to move 9.3 into production?
JHall are you joking? It’s our responsibility to test the software for ESRI with all its resources? The user community can help, but relying on them to do the work for ESRI is weak. Its called QA, I have to do it for my customers and ESRI should do a better job for us.
Not that QA should come with software that costs $40,000 or anything…
10
jxn
// Jun 26, 2008 at 7:12 pm
Or documentation…
11
j.geeraerts
// Jun 27, 2008 at 1:17 am
damn! in U.S. ArcView costs 1.500 dollars, but in Italy its price will be about 3.000 euros (VAT included) …
12
Shrek
// Jun 27, 2008 at 10:06 am
From the perspective of the user of Manifold System GIS 8.x, it is interesting to see what the new release of ArcGIS 9.3 does not include, rather than what it does. Namely:
No support for 64-bit Windows. So ArcGIS users still cannot use more than 3.5Gb of RAM accessible to 32-bit Windoes OS’s
No support for massively parrallel, potentially teraflop-level processing capabilities provided via NVidia GP-GPU cards + CUDA, or the equivalent using Open-CL with AMD/ATI GP-GPU cards.
From this one would be forgiven for thinking Arc was an abbreviation for archaic. On this basis alone the product sucks. I pity anyone who will have to use (or endure) it.
13
MicahWilli
// Jun 27, 2008 at 10:32 am
Shrek: ouch ease off the hate a little. There’s support for 64 bit systems, the software just doesn’t fully take advantage of 64bit and multi-threading.
jxn: Lets talk about that a little. I’m just about to do a fresh install of ArcGIS Server Enterprise Advanced. Should I do 9.3 & just hope it all works or put on 9.2 SP5 & wait for the dust to settle?
This is an honest question.
14
jxn
// Jun 27, 2008 at 10:50 am
Micah-
I dont have the answer - maybe it depends on how much you are depending on your own code vs the esri templates? Are you starting from scratch or do you have code to migrate? What are you working on?
Can anyone here say for sure that 9.3 is backwards compatible? Im not sure how the MS Ajax library vs client callbacks switch affects things. Im prob not the person to answer this as I dont have any AGS work (other than SDE) at the moment so I havent spent the time doing research and testing.
When 9.2 AGS Ent Adv SP2 or so was first released I had to create an editing application using the editortask as a starting point, but there was no documentation so it took a really long time. But it also doesnt really make sense though to build something that uses out of date software unless it can be upgraded easily….
Perhaps im overly cautious, we still havent implemented SP5. Im generally of the opinion that if it aint broke, dont fix it, and if its new, wait a bit till the obvious kinks are worked out.
15
kj
// Jun 27, 2008 at 10:56 am
Hey Shreck - what about all the other stuff that mfold continues to not do at each release that ArcGIS does, like - free software (arcgis explorer, arcgis javascript APIs, etc), GIS installed on lots of devices and different OS’s like linux, PDAs, cell phones (not just windows desktop and web browsers), and so on. Maybe we should go post all that stuff in your forum: Http://forum.manifold.net/forum/t67219.1
16
Chris C.
// Jun 27, 2008 at 11:55 am
damn! Manifold $295 USD in the US
$295 USD everywhere else.
You’re right kj Manifold should be giving us users free software to appease us
17
MicahWilli
// Jun 27, 2008 at 11:59 am
Jxn
As far as AGS goes (the web ADF) I’m starting from scratch.
But We are already heavy users with IMS and SDE. and that’s what I’m a little concerned with. I will be installing on a non-production (new) server, then making the switch. so there’s as little downtime as possible. The obvious benefit to installing 9.3 now is not having to do it again in 2 months when that server will be production, but at least then it will be seasoned a little. I will know what I’m getting into.
There’s nothing wrong with 9.2 but I always like the newest, i.e. I’m currently running vista
That being said, I’m leaning your way…caution. My machine is one thing, our county-wide data server is another matter.
I think I’ve made my decision.
18
abridge
// Jun 27, 2008 at 2:19 pm
See http://forum.manifold.net/forum/t67219#67230 for Manifold user group discussion. While most of us Manifolders love the product, we don’t all need to come down hard on ESRI (aka apology for rudeness). Having said that, if I were a ESRI user I would be asking hard questions to them regarding their strategic vision in light of big advances in computing.
19
Dimitri
// Jun 27, 2008 at 3:56 pm
Look, if someone wants to praise ArcGIS 9.3 that’s OK, but let’s not have any nonsense about it being in any way a 64-bit thing. Praise it for other things but not that.
MicahWilli, the above is seriously misleading based on what ESRI says.
ESRI’s “What’s New in ArcGIS 9.3″ page provides the following FAQ answer:
64-bit Windows can execute almost all legacy 32-bit applications unmodified but it does so exclusively in 32-bit mode. When you run a 32-bit application like ArcGIS 9.3 in a 64-bit Windows O/S, the moment that 32-bit application launches you have just lobotomized your 64-bit Windows system back into 32-bit mode.
Contrary to your statement’s “fully” word, which implies there are some advantages taken by ArcGIS 9.3 in a 64-bit OS, you do not get the benefits of 64-bit Windows because 32-bit ArcGIS 9.3 is forcing Windows to execute in exclusively 32-bit mode.
In fact, the situation is worse because you have all the limitations of 32-bit operation (less reliability, lower performance, painfully small process space, etc.) together with the additional uncertainty of running within an emulated 32-bit environment.
The only thing that ESRI’s “supported” statement means is that 64-bit Windows can execute in 32-bit mode a legacy 32-bit application like ArcGIS 9.3 and that ESRI’s tech support department has verified that. This may be a tribute to Microsoft’s ability to provide emulators that can execute old-fashioned code in modern environments, but it is nothing at all for ESRI to be proud of, especially given that “no immediate plans” bit.
By the way, note that the ESRI answer above comes in response to the FAQ question:
I’m surprised there is not a word on Windows Server 2008, since both XP x64 and Vista x64 are very old news. By far the best 64-bit operating system out there is Windows Server 2008 x64, recently released by Microsoft and supported from inception by the likes of, well, Manifold. How can a mid-summer 2008 product fail to support the number one Microsoft operating system released in early 2008?
If you don’t think that’s a big deal, ask any Vista guy who has tried Server 2008 why he instantly switched to Server 2008. “It is the operating sysem we all wished Microsoft had done instead of Vista,” I heard one engineer say. It is so outstanding a Windows, the best ever, that people are using it as a desktop OS and many (including me) are installing it on their notebook computers as well. Not supporting Server 2008 is a big deal.
By the way, I agree that Shrek’s last line about pity was harsh. But in all fairness to Shrek it is truly astonishing, first, that a very expensive, supposedly major new product released in the middle of 2008 does *not* provide true 64-bit operation and, second, that the vendor blithely announces no immediate plans to work on 64-bit product. Amazing, especially when you consider that since this is the top FAQ on their “What’s New” page for 9.3 they’ve obviously had a lot of questions on true 64-bit capability.
Upgrading 32-bit code to 64-bit code is not at all difficult - it is straightforward work any journeyman programmer can do. But it is labor intensive work that requires a commitment from management to back with adequate resources. It’s amazing that, what? four years? after 64-bit Windows was first available ESRI is still sitting on its hands on this. If they are taking this long just for the baby-step of upgrading to true 64-bit code in their flagship product, it will be 100 years before they ever do something truly difficult like CUDA.
By the way, to see CUDA in action reducing a job that takes one minute without CUDA to a mere 2 seconds with CUDA, there’s a live demo in the video presentation at
http://www.manifold.net/video/Supercomputer_GIS.wmv
Unless you are happy if your work takes 30 times longer than it needs to take, then CUDA is nothing to disregard. It, too, is a big deal and the first time you do in a few seconds what used to take many minutes you too will understand why Shrek allowed his scorn to show, albeit in an overly harsh way.
Despite all that, congratulations to ESRI for issuing 9.3 - not easy getting those big products out the door as every developer knows.
20
James Fee
// Jun 27, 2008 at 5:09 pm
Dimitri: There is 64-bit support for Server 2008:
http://wikis.esri.com/wiki/display/ag93bsr/ArcSDE+Microsoft+SQL+Server+Database+Requirements#ArcSDEMicrosoftSQLServerDatabaseRequirements-SupportedOperatingSystems
That is for ArcSDE, but if you check out all the server products, you’ll see 2008 support, but only ArcSDE is really a 64-bit application.
21
JeffB
// Jun 27, 2008 at 6:55 pm
I’ve read that ESRI intended to release full 64-bit versions of more products, but in their efforts to get the software out to the users, they cut some of the 64-bit builds. They will release them at 9.3 SP1. I can’t remember if that was SDE only (on Windows and Linux systems that currently only support 32-bit), or if that included Desktop. If I find the article again, I will post a link to it.
22
Dimitri
// Jun 28, 2008 at 10:52 am
ESRI provides an authoritative answer to this in their what’s new for 9.3 page (see http://esri.com/software/arcgis/about/whats-new.html ) is . Their full answer to the very first FAQ reads:
Having 64-bit ArcSDE is darned near useless to people who are doing GIS on the desktop. If you are not running 64-bit software on your desktop you don’t get 64-bit reliability, 64-bit performance and 64-bit capacity on your desktop.
ArcSDE is just a file cabinet. It is not the software you actually use for significant desktop GIS tasks like editing, rapid visualization, spatial overlays and the like. It does you no good to have a file cabinet with a normal IQ if the tools you actually use to do anything with the information are slow, unreliable and stupid.
By the way, the reason ESRI made SDE 64-bit was because they were getting tossed out by the DBMS guys. DBMS administrators tend to be a very technically astute group and so they know that 64-bit operation has been a “must have” now for years. No way is any sensible DBA who is already running 64-bit SQL Server or Oracle on Server 2008 going to trash the IQ of his servers by running moronic 32-bit code on it.
ESRI’s quote makes it clear: they are doing 64-bit SDE but have no plans to do so for the desktop.
The interesting thing about all this, by the way, is that 64-bit SDE is a “don’t care” because it’s a terribly inefficient idea to interpose middleware like SDE between a modern, well-written DBMS and a modern, well-written client. There’s no point to SDE whatsoever except to put a leash around your neck and hand it to ESRI. Anyone who really cares about performance (the usual 64-bit constituency) isn’t going to be using SDE if they can help it.
A modern, well-written client like Manifold will not only be 64-bit already, it will communicate directly to the spatial DBMS using that DBMS’s own native protocols (like OCI, Oracle’s native interface) using the DBMS’s native geometry types (like SDO_GEOMETRY and GEORASTER for Oracle). It will connect with lightning speed directly to that DBMS using massively multithreaded access and multicore access as well if you have multiple cores on your desktop (all modern desktops and most notebooks do). Obsolete architectural notions like SDE just get in the way of that high performance relationship between modern super-client and modern, spatial, super-DBMS.
From a customer efficiency and performance point of view, instead of wasting time band-aiding a technology that already is obsolete (SDE), it would have been far wiser for ESRI to have invested those resources into a 64-bit / parallelized modernization of client software like ArcGIS Desktop while providing direct connectivity to native spatial DBMS. They didn’t do that because they know that at the end of that process (which would have taken them much longer because it is more difficult to do than an upgrade to the relatively limited black box interfaces of SDE) they would have faced competition on the desktop already far ahead of them and charging 1/40th of what they do.
I realize that was not a very appealing scenario, but frankly if I was ESRI I’d rather take my chances competing with Manifold on the desktop than continue to compete with the major DBMS companies like SQL Server and Oracle by promoting SDE in opposition to the DBMS vendors’ native standards. And no matter how much people will play nice for the cameras, if you know your technology you know that given what Oracle and Microsoft advocate for their spatial DBMSs there is no future for alternative proprietarizations like SDE. It is only a matter of time and I would not bet on ESRI to win against both Oracle and Microsoft.
By the way, multithreaded is not the same as multicore. It makes sense even in a single core system to be multithreaded. To take an example, if you have a fast desktop even with a single core it will probably operate faster than even a fast DBMS server can service it through even a fast network link. Running multithreaded on a single processor you don’t have to do everything linearly, like waiting for a data access to come back from the DBMS. If you are multithreaded, as soon as one thread starts bringing in data from the DBMS you can start working with what you have, such as rendering it, in a different thread while the first thread continues to fetch data. You can see this in Manifold with things like progressive rendering where some threads are rendering even as other threads continue to fetch more data from the data source. Of course if you have multiple cores to use that’s so much the better.
Guys, don’t fall for this “oh, well, it is 64-bit because SDE is 64-bit even if nothing else is.” That’s smoke and mirrors aimed at people who are dumb enough to get suckered back into time-sharing. Time-sharing got killed off by distributed computing for a very good reason that is even more true today: 100 processors serving one user are way faster than 100 users time-sharing one processor.
23
James Fee
// Jun 28, 2008 at 10:58 am
Dimitri: For some of us, we only run 64-bit on servers. I have no plans to move to 64-bit on workstations anytime soon.
I’m sure there are people who need it, but the only thing I’m concerned for is 64-bit on servers and finally I can run my whole spatial database stack on 64-bit.
If I used Manifold, it would be on a 32-bit system so at least for me desktop support is irrelevant.
24
j
// Jun 28, 2008 at 2:24 pm
@ James Fee
on the flip side - both ourselves and a significant chunck (not the majority, but a good size - maybe 15%) have been waiting to switch to 64bit on the desktop side but have been waiting for more 64bit app’s to make it worthwhile and avoid mixing environments. From what I’ve heard from other groups for some of our office app’s, they’re starting to move towards 64bit so we’re setting up a few 64bit “utility” stations. Early testing has shown that some operations that took around a week from start to finish are now down to a couple of days, and we anticipate that we can get some of that down to a day or less after we get some of our processes set up better.
In the end - I do think that everything will eventually go 64bit, whether by Microsoft or someone else. It just seems like a logical progression and evolution. Therefore, I would think that the faster that any application provider (GIS or otherwise) gets themselves positioned to take advantage of this, I think it would be probably for the better.
25
James Fee
// Jun 28, 2008 at 2:49 pm
j: I totally understand and believe ESRI should be supporting 64-bit on the desktop in 2008. Don’t get my comment wrong.
I don’t think anyone here is blinded by excuses as to why there is no 64-bit support for desktop. My organization has no plans to move to Vista and no plans for 64-bit desktop support. In that environment I could care less about anyone supporting 64-bit Vista desktops.
YMMV of course.
26
Jesus
// Jun 28, 2008 at 2:54 pm
“I was ESRI I’d rather take my chances competing with Manifold on the desktop“
Dimitri, even you crack me up these days. Keep drinking my friend!
27
j
// Jun 28, 2008 at 7:01 pm
@ james Fee
Fair enough
Cheers.
28
Dux
// Jun 29, 2008 at 11:15 am
*chuckle* Sometimes you just gotta get a chuckle out of life. You gotta give credit to Dimiti and his loyal band of Manifold users. Outside of the blogsphere, you hardly know this company exists. I do appreciate these threads and know for a fact that almost anytime the word “ESRI” or “Arc” hits the pages of this blog, the term “Legacy” will certainly be following within the day.
29
Chris M
// Jun 30, 2008 at 7:39 am
@James Fee
Excellent point about usefulness of 64 bit desktop software. I don’t see many users rushing to switch to 64bit XP or Vista any time soon. I am personally seriously debating delaying desktop replacements scheduled for next year until after the new MS OS is available or at least Vista SP 2 is ready and thoroughly tested. In general I think most mainstream software vendors have dropped the ball on port applications to 64 bit. I guess they all thought it would take longer for this technology to make it for the desktop and probably would have if the AMD vs INTEL war did heat up like it did.
30
Dimitri
// Jun 30, 2008 at 8:50 am
—————
Jesus, Dux,
Whoa! Dinosaurs bellowing!
Hmmm… hit a nerve did we? Interesting how quickly people want to change the subject from ArcGIS 9.3 to something else when the antique internals of ArcGIS 9.3 come up. Let’s try to keep this on ArcGIS 9.3 and on the implications of that.
I’d like to hear your opinion of the substance of my remark, that by taking SDE 64-bit and failing to do so for ArcGIS desktop, the new 9.3 shipment shows a misguided emphasis on what is ultimately a doomed attempt to compete simultaneously with Oracle, Microsoft, IBM, SUN (MySQL, you know) and the open source community (PostGIS) by trying to push SDE in competition with native spatial DBMS standards.
ESRI users are already paying thousands of dollars per license and are asked to contribute thousands more in maintenance fees, so if one of the messages of ArcGIS 9.3 is that those funds are going to be burned up on technology that ultimately is a lost cause that’s something folks who care about their hard-earned dollars usually want to know. Especially in tough economic times the wise GIS professional doesn’t buy stuff the vendor tells them is not being evolved. ESRI today tells you that ArcGIS Desktop is not being evolved into 64-bit capability, “no immediate plans” and all that, but that SDE is. No one who cares about getting the most value and reliability out of their desktop GIS dollar should ignore that bit of 9.3 news.
It is your money, the thousands you spend per license and the thousands more in maintenance, that finances the tilting at windmills SDE thing. So when ESRI spends your money in that direction and fails to spend it on making ArcGIS 9.3 a modern and reliable product on the desktop or fails to spend it on making ArcGIS 9.3 Server anything but an utter dog for performance (so that it might be able to meet modern expectations for performance on the desktop despite being a return to time-sharing), it is your money they are burning up trying to push SDE upwind against the combined force of Oracle, Microsoft, IBM, SUN and the open source DBMSs.
I suggested that instead of spending your money on an ultimately-doomed promotion of SDE (taking it 64-bit, etc.), it would have been wiser to modernize ArcGIS for competition on the desktop, whether you think that will come from a modern ArcGIS Desktop capability or whether you believe that will come if the performance problems with ArcGIS Server can be fixed. Since you mocked my suggestion that it would be easier to compete with Manifold on the desktop than it would be to push uphill against the combined weight of Oracle, Microsoft, IBM, SUN and the open source community, perhaps you would share with us why you believe my comments weren’t true, and that it would be easier for ESRI to compete against that massive crowd?
James,
If you don’t have 64-bit desktop software you don’t have a “whole spatial database stack” running 64-bits. The chain is only as strong as its weakest link so having 64-bit spatial DBMS connected to 32-bit desktop software is like taking a modern eight-lane highway and connecting it to a narrow country road: massive slowdown and traffic jam.
If, on the other hand, what you are saying is that you don’t care about desktop software and you have bought into the notion that we are all going to return to time-sharing via exclusively server-side software, well, that merits a thread all on its own.
The majority of people doing GIS have to wrangle with it on the desktop day-in and day-out putting in their eight hours a day and change doing stuff like editing parcel maps, generating maps for various county departments to use, responding to emergency demands, computing sales tax with complex spatial SQL, doing topology overlays to see how their utility rights of way intersect special legislative regions like wetlands or tribal lands or whatever. I could be wrong about this but I bet most readers of this blog do that sort of thing and are darned well interested in whether or not a return to time-sharing is going to give them the increased performance they need to deal with the increasing demands of their jobs at a time of reduced funding.
Time sharing has been discredited over and over because of the simple fact that one processor per user gives you much faster performance than 100 users per processor. In modern times it’s even more skewed, because now we can say that 100 processors per user gives you much faster performance than 100 users per processor. You can see how that is true today more than ever by seeing how Manifold can use CUDA (256 processors per user) to do in 2 seconds what takes 64-bit deskitop software a minute and what time-shared software like ArcGIS Server can’t do at all, not even in hours. [For the new link to the short CUDA demo, see the news page at http://www.manifold.net/info/news.shtml
James, I respect that for a strictly local, atypical task that is not representative of what the majority of GIS users do every day with their eight hours and change on the job you may be willing to accept a time sharing focus that is exclusively server-side 64-bit SDE. But I don’t think very many other people will accept that and offering an implied excuse for the failure of ArcGIS 9.3 to modernize by discussing 64-bit operation in such limited terms does not address the substance of the matter.
Going 64-bit on the desktop is a *huge* deal. Everyone reading this blog who has suffered through crashes and slow performance from ESRI products could get rid of a lot of that by simply going 64-bit. I suppose somewhere there may be someone who is happy about having their time wasted through endless crashes, but most folks realize that such problems are an especially egregious waste of time and money, especially so since they could be so easily eliminated in large part by going 64-bit.
And, contrary to the implication of…
… Vista is not the same as 64-bits.
There are four 64-bit Microsoft Windows operating systems now in common use for over five years: Windows Server 2003 x64, Windows XP x64, Windows Vista x64 and now Windows Server 2008. Whatever you think of Vista, at least three of these are the best Microsoft Windows releases of their era, with Windows Server 2008 x64 being the best Windows ever, whether you use it server side or on the desktop as many people are starting to do. The only thing that matters is making sure your GIS vendor is not some engineering cripple who finds it difficult to support whatever 64-bit Windows you choose.
I realize some people still use Windows 98 and some people never will modernize. But I presume that’s not the focal point of interest for people who care enough about the efficient use of new technology to want to read about ArcGIS 9.3. So let’s not have any of this “Oh, I don’t care about modern stuff” talk as an excuse for the failures of 9.3 because if that’s what’s going on you may as well skip 9.3 and continue talking about ArcView 3.x.
Last, but not least, while I suppose I should be mature and not respond to an ill-considered taunt, I cannot resist commenting on…
Dux, I realize you intended this as a put-down of Manifold, but it really is more embarrassing to you because it indicates you don’t travel in circles that are aware of contemporary technology.
Everyone reading this blog realizes that within the tens or even hundreds of millions of people who have moved to 64-bit systems there is some significant number of GIS users who have moved to 64-bit Windows as well and who are now running 64-bit GIS applications on their desktops.
To state what should be painfully obvious, every one of those 64-bit GIS users is *not* using ESRI, since ESRI has no 64-bit GIS products for those users. Instead, almost all of them use Manifold, since only Manifold delivers a full stack of full-featured professional and enterprise GIS product in true 64-bit form. In 64-bit use it is all Manifold and zero ESRI. If you believe what ESRI says, that will continue to be the case for the foreseeable future.
Therefore, Dux, when you state you “hardly know” this company exists, you are telling us you don’t do 64-bit GIS and neither do the people you know. You are also telling us your usage is so undemanding that neither you nor the circles you travel in are exploring 64-bit operation, because if you were you’d be studying the 64-bit GIS product, Manifold, that outsells all others over 100 to 1 in 64-bit markets.
I’d be the first to agree that a tit for tat debate about who is the big fish in which small pond can get pointless really fast. But there is some merit to the notion that you shouldn’t celebrate being the big fish in a small pond that’s drying up. That would be ESRI in the 32-bit pond. If you think 64-bit GIS in Windows is an even smaller pond. too small a pond for anyone to care about, fine. You’re entitled to your opinion. But however small or large the 64-bit GIS pond may be, we are the big fish with years of experience in that pond, not ESRI.
Further, not only is it our pond that is growing, it is our pond that attracts the technology elite, because no one who really has it going on in technology will tolerate the stifling limitations and stupidity of lobotimizing their wonderful 64-bit processors into 32-bit emulation mode. Just won’t do it. You know who you are!
It is now five years since Windows Server 2003 x64 brought 64-bit Windows out of the development community and onto millions of computer systems worldwide. The fourth generation Windows Server 2008 x64 has emerged as the undisputed Best Windows Ever in the opinion of virtually everyone who has experienced it. Developing complex software products is not so easy that a company can afford to waste five years or more and then catch up, yet even at this very late date the message of ArcGIS 9.3 is that ESRI has no intention of catching up on the desktop to what has emerged as the number one trend in mainstream computing for professional applications: 64-bit operation.
31
KoS
// Jun 30, 2008 at 9:50 am
Dimitri…You might what get your head out of your ass sometimes. I know you are a company man and have to spout what you spout. Myself and others keep telling you, your tack doesn’t help your cause.
Manifold isn’t used by alot of people as compared to the other vendors(overall), period end of story. Not many people in the field uses your software, especially in my part of the world. I can only think of two business in the area who are using y’alls software(as their main package).
The market share numbers I keep seeing shows it to be true. Y’all have a small group of users in the private sector. Where as in the public sector, I would be willing to bet, close to zero percent. Especially after pulling the software out of IU. At IU you had the biggest user base of Manifold in the state.
Regardless of the whys and what nots. Y’alls software, numbers wise, lags behind. Y’all may be the greatest thing since sliced beard. The numbers sure don’t show it. If y’alls software is that great, why are not more people beating a path to your doorstep?
O’wait, I know why. People are just too stupid to see a great thing!
KoS
ps. yes I still bad mouth your company, even if it’s the best solution for a given task(s). And that is exactly what I tell people. It’s cheap and can do what you need, very well. But don’t buy it, all due to one man’s mouth.
32
abridge
// Jun 30, 2008 at 10:03 am
KoS,
Here in British Columbia the provincial government is a bastion of ESRI. However, the geological survey group has recently entirely switched to Manifold, and now deliver their industry standard geological mapping products in Manifold format e.g. http://www.em.gov.bc.ca/Mining/GeolSurv/Publications/GeoFiles/Gf2008-1/toc.htm (which by the way works really swell, as the one file .map format gives you all the data in full format as they intended). They have even written custom geology symbology available on their website.
I corresponded with their lead guy about how they made the decision to switch, and he said they have equiped their entire crew of field geologists (10+) with laptops and Manifold for less than the cost of 1 year of top tier ESRI *support fees*. Also, as mentioned, the .map format is ideal for distributing a full set of geology, roads, lakes, streams etc. data for a mapsheet in one file.
Another example, a friend who is a habitat supply modeler in the government just ordered Manifold on my advice as he is sick of having to access ESRI through the server sharing system that the government has implemented. Just wait till he discovers CUDA based raster improvements. (Think LIDAR DEMS and processing speed).
It is happening. It might take another 5 years for full penetration, but I see it coming.
33
KoS
// Jun 30, 2008 at 12:49 pm
Abridge….good for y’all! In all seriousness. If you are happy, that is all that matters.
As for govt use here. Manifold won’t be installed on any USDA machine any time in the near future. And if I was a betting man, the state govt won’t be using it anytime soon either.
I guess too, y’all should thank the American taxpayer for this gift.
A cheaper GIS system which is better than the rest. Which ironically, more than likely, is in more use outside of the US compared to within.
If I believe the history, Manifold started out as a DoD project.
As far as market penetration. In our neck of the woods. Manifold needs to do a better job of marketing themselves, otherwise, they will still lag behind in sales and use. Plus, it doesn’t help to have someone who speaks for the company, basically tells people they are stupid and don’t know their ass from a hole in the ground. Just because they haven’t seen the light of Manifold.
KoS
34
JW
// Jun 30, 2008 at 1:07 pm
Dimitri, you are really blowing a lot of smoke. It must be nice to be in the position you are in, using a GIS that was built from the ground up just recently, as opposed to years of code meant to support customers legacy apps. I wish all of us could throw away all of our old code and start from scratch. Then we could almost hold our noses as high as yours.
Yo uare right about one thing though, you guys do attract the technology elite(ists). Probably also have the zealot user base covered to. If you all would just make an OS X version you could tap into an instant user base of elitist zealots.
35
j
// Jun 30, 2008 at 3:51 pm
What the heck -
@ KoS
I would suggest that your comment about Manifold not being used by “a lot” of people all depends on what circles you travel in. We keep running into more people/groups/clients using it, and the majority of those are not private but public agencies such as municipalities and authorities, as well as oil and gas which aren’t exactly “small”.
In the interest of being open and fair - not everyone that I know who uses it use it as their “main” GIS. Some use it more as a companion application. Hey - they’ve speant $1000’s on their “main” GIS so an extra few $100 isn’t going to worry anyone. This, and the fact that some people don’t use it for GIS at all but as a database interface have lead to a good number of users in our circles.
Personally, I think Dimitri’s commet re. a “big fish in a small pond that’s drying up” may not be too far from the truth. As GIS becomes less of a niche and more of a broad field, I think that applications that are less of a “this is a GIS” and more of a “this is a database app. that can do spatial stuff” will start to grow. This also means that groups who’s focus isn’t on GIS explicitly will be looking for less expensive applications that make better use of harware and software. I think that this probably helps account for some of the growth I’ve seen from those using qGIS.
You may be right though - this is all from what I see north of the border in “my” circles. More then a few of our conservation authorities use Manifold (only one using ESRI in our area that I know of) and it does appear to be growing.
@JW
Well - Manifold the company has always been pretty good at responding to user requests, so if enough Mac users say they want it to make it worth their while … who knows;-)
36
JW
// Jul 1, 2008 at 5:08 am
well, there is your problem right there, you are in Canada. Every Canadian GIS person I’ve ever met is always tinkering with 5 different packages and never has any substantial analysis from any.
37
ChrisW
// Jul 1, 2008 at 7:54 am
Er…can I ask a dumb question?
Dmitri says “it’s a terribly inefficient idea to interpose middleware like SDE between a modern, well-written DBMS and a modern, well-written client”. As a DB person new to GIS, my initial (ignorant) impression is that he may be right. I can see why people would choose ArcGIS for the desktop (despite the fact it crashes all the time and still doesn’t play nice with Vista), or even IMS for web mapping, but what on earth is it about SDE that people have been willing to pay such massive licence fees for (on top of the SQL Server licence for a non-spatial DB that usually seems to go with it)?
I know ArcGIS Server is supposed to move on from SDE, and I’m not trying to stir up more Manifold wars either, just genuinely curious as a GIS newbie about the perceived benefits within the GIS industry of what looks to me like an expensive and largely superfluous product.
Cheers, Chris
P.S. @JW - didn’t the Canadians pretty much invent GIS in the 1960s, or is that just Canadian propaganda?
38
DannyV
// Jul 1, 2008 at 9:43 am
@ChrisW: That is just what those Canadians want you to believe.
Ah yes, every ESRI post here turns into Manifold is better discussion. You’d think by now they’d just have their own blog. Oh well, at least we have the cats.
39
JW
// Jul 1, 2008 at 10:22 am
I’m wondering exactly how many people could actually take advantage if all that 64-bit goodness we keep hearing about. I have yet to encounter any. I would like to think I’d have a need to process 8Gb raster sets in RAM, but the reality is I never come across that need.
I guess I’m not tech 1337 enough.
40
James Fee
// Jul 1, 2008 at 10:50 am
JW: Think of this logically as Manifold does:
1. GIS is a small pond
2. 64-bit users of desktop XP or Vista is a small island of the Windows user base
Thus manifold’s target is a small island on a small pond. This doesn’t mean that 64-bit support is not needed (I clearly understand it is), but to claim that it is anything more than a small clique compared to enterprise windows users as a whole is misleading.
Those who need 64-bit desktop support have options and it is good to know Manifold is at the forefront of this clique, but of all the clients I have and all the users I interact with daily, ZERO have 64-bit XP or Vista.
Call me unconvinced it is even an issue on the desktop.
41
cluster
// Jul 1, 2008 at 10:55 am
Hi Chris,
My 2 cents:
If we treat GIS as a system, as a business component, then one of the most important thing is support, in terms of funcationality and service.
Why not using mysql/postgresql but pay the hefty fee to Oracle/SQL Server/DB2?
Functionality:
ArcSDE does provide versioned editing/archiving/
topology/geometric network/annotation (some of them are also available in Oracle, but you can try any serious client software, e.g. from Leica geosystems and draw your own conclusion).
Oracle (up to 11.1.0.6) hard-coded sdo_ordinate_array as varray(10485746), which limits the number of points user can store in one SDO_GEOMETRY object.
SQL Server: please read sql server 2008 document on the spatial data, page by page, you will see it still has a long way to go.
Performance:
No one likes to talk about performance in database world (or everyone likes to talk since everyone is number 1). But you can try it yourself, compare SDO_GEOMETRY, SDELOB, ST_GEOMETRY, …, etc.
Client:
Again, you will have to try all different client software yourself to draw your own conclusion.
If you have limited usage, you will find you really have so many options.
If you need a complete GIS system, well, again, try it yourself.
42
scooter
// Jul 1, 2008 at 11:15 am
cluster, perhaps that’s just me but I don’t see how what you say calls for middleware such as SDE. Assuming Manifold is as good as Dimitri paints it to be, is there a reason not to do any of the things you mention either on the DB or on the client as he suggests? I can’t see any.
43
scooter
// Jul 1, 2008 at 12:06 pm
BTW, James, one reason to go 64-bit on the desktop has been mentioned on your blog:
http://www.spatiallyadjusted.com/2008/05/08/developing-in-a-virtual-environment/
Virtual machines.
44
Dimitri
// Jul 1, 2008 at 12:18 pm
Dannyv,
For the record, I’ve been discussing ArcGIS 9.3, specifically stating that it would be better for ESRI to invest more resources into modernizing ArcGIS 9.3 Desktop than pushing resources into what I’ve suggested is a dead end, SDE. I joined this thread to correct the wrong implications in a post that ArcGIS 9.3 is “supported” by ESRI in 64-bit Windows.
Except for James making the case that 64-bit operation is a “don’t care,” in response, I’ve mostly gotten “ad hominem” flames from touchy members of the ESRI community who drag Manifold into the conversation instead of addressing the substance of what I wrote, about whether this neglect of the desktop is really a good course.
If you go back to my postings you’ll see that at no place did I say “Manifold is better.” If anything, I wrote an implied put-down of Manifold by stating it would be much easier for ESRI to compete with Manifold on the desktop than to compete in spatial DBMS against giant companies like Oracle and Microsoft. For that matter, even in 64-bit operation I’ve simply stated that ESRI has decided not to fight in that arena. I’ve not said anything about if they decided to join the fight that their product would be better or worse. We don’t know because they haven’t gone 64-bit and have no plans to do so.
It is true I have used Manifold as a reality check from time to time to show that some things are not only possible, they are highly conventional and indeed are being done by others.
It is useful to have such a reality check on the discussion so that there is a metric as to the reasonableness of different options when talking about what the changed focus in 9.3 away from the desktop and onto SDE and server means to GIS . Manifold is a very useful counter-example to some of the excuses one hears for why 9.3 isn’t getting modernized.
For example, there are folks who will tell you that the difficulty or extra cost of going 64-bit is somehow a good reason why ESRI has ignored this for five years.
I say “nonsense,” and can point to an example. If Manifold can do it years ago for a product that simultaneously combines what is in ArcGIS Server, ArcSDE and ArcGIS Desktop and lots more in addition than clearly this is not something that should be beyond ESRI’s ability.
Likewise, one hears all sorts of guff about things like SDE, yet seeing in reality that such middleware is not necessary by the example of a modern client connecting directly to a modern spatial DBMS proves you don’t need no proprietary middleware for that.
To take a third example, when you hear people trying to talk you into a return to time-sharing with ArcGIS Server and I tell you “10o processors per user is much faster than 100 users per processor” it is useful to show by actual example that this is not opinion, it is demonstrable fact. The CUDA demo with Manifold shows a real-life example. Yes, that freaks out the dinosaurs who would rather not have such an emphatic, inarguable demonstration that time sharing is still a bad idea, but, well, too bad for them. A concrete demonstration proves the point in a way no amount of opinion can do.
If you want to choose some other GIS program to use as a counter-example that proves my point, fine. Use what you want.
I primarily use Manifold for two reasons: first, because it makes an especially sharp example to dinosaurs who think of Manifold as a much smaller company. If what they say is true, than my point is even more emphatically made because if a very much smaller company can do all this stuff then it is an especially telling lapse for ESRI not to.
The second reason is I know Manifold so I can pick out accurate examples more easily. Don’t have all day to write this stuff, you know.
By the way, what surprises me about these ESRI blogs is the degree to which truly massive shifts in focus go unchallenged. With ArcGIS 9.3 ESRI is telling desktop users to get ready to jump back into time-sharing via ArcGIS Server. That’s a huge deal, yet very little discussion. Suprising. We do have our forum at Manifold (http://forum.manifold.net) and I am very proud to say that the Manifold user community is highly demanding and doesn’t hesitate to debate emphatically and to criticize the company. That’s how the product improves. But here in the ArcGIS thread we’re talking about a trend, a shift in focus away from the desktop, a taking away from resources that could be used for modernization and progress on the desktop, that is of general interest to many in the GIS and computing communities.
45
ChrisW
// Jul 1, 2008 at 12:52 pm
@cluster: Thanks for that.
Food for thought, although I’m less convinced by the non-spatial arguments about DB performance/support, as I reckon these might well be addressed more cheaply in other ways. But Arc Server still seems to have some GIS-specific advantages, from what you say.
I think Dmitri disagrees, though!
46
Dimitri
// Jul 1, 2008 at 1:03 pm
James,
Some clarification:
We believe legacy GIS is a small pond. Mainstream GIS is very much larger.
Yes, GIS is indeed a small pond if you price it at thousands of dollars per seat, sell it using legacy sales and marketing techniques and you fail to leverage mainstream standards. It is a very much larger market if you make it affordable to the masses at $245 a seat, make it accessible to the masses by direct sales through Internet and you support broad mainstream uses by embracing mainstream standards. You then can achieve user bases of many tens of millions of users.
That’s not as big a market as the entire host Microsoft ecosystem that numbers in the billions, but it is far beyond small pond status like legacy GIS markets now have.
I also disagree that desktop use of 64-bit Windows is a small island. It is less than 50% of Windows to be sure, but it is very large in numbers already driven by a variety of factors: gaming, graphics arts and a wide variety of professional programs that have gone 64-bit. And it is getting very much bigger.
I think you’re getting a skewed picture of this that leads you to exactly the opposite conclusion of what the trend is. If the content of your blog is any guide (I’m not criticizing, I’m just noting a highly evident focus), you hang around in ESRI circles. Well, OK, that’s a self-selection process in that because there is no 64-bit desktop GIS in ArcLand it is expected you won’t know anyone who does 64-bit desktop GIS.
Back in the mainstream the situation is different. Not only are there very many mainstream folks eagerly discovering GIS, those folks tend to be early adopters, technology elite, inquisitive types and, frankly, very high technology types. Running 64-bit is no big deal for them. In fact, they tend to wonder what kind of backward person would *fail* to run 64-bits and thus fail to take full advantage of the 64-bit hardware and 64-bit memory they’ve already paid for. They *expect* that for any serious application you’ll use 64-bits.
Another factor is that the “New GIS” users tend to over-use image servers. A lot of traditional users have had the experience of running into someone new to GIS who uses images for everything instead of using vector data sets. That’s true in New GIS as well, so they have a tendency to plug an image server into Manifold and then start downloading a few gigabytes of iamges that they use as convenient background maps. Those folks are instant candidates to go 64-bit.
For the above reasons, while in legacy GIS you don’t see 64-bit desktop GIS (it simply doesn’t exist) in the mainstream not only is it numerically big, it is actually a much larger share pro rata because people a) are more used to the idea and b) they tend to operate GIS in a way that uses big images and thus tends to give an extra incentive to 64-bit operation.
Ah… one last thing: I’ve never suggested that 64-bit desktop GIS usage in Windows is “more than a small clique compared to enterprise windows users as a whole.”
First off, enterprise windows users as a whole are strongly moving to 64-bit Windows, so any discussion of whether 64-bit Windows is happening for that crowd is moot. There’s a reason, you know, why even laptops sold to Enterprise customers (like ThinkPads) are sold with x64 Windows installed as an option.
Second, my point was not comparing 64-bit desktop GIS users to enterprise Windows users as a whole. I simply stated that numerically there are a lot of 64-bit Windows users and that quite a few of them are doing 64-bit GIS and that none of them are now using ESRI. It’s a self evident point.
In point of fact, even though the number of x64 Windows users is a subset of all Windows users and those who are doing GIS is a subset of that, the number of Windows users is so large (what? over a billion by now?) that even that subset of a subset is a very large number that dwarfs anything in legacy GIS. We are just starting to cut into that market, of course, but for now we have it all to ourselves.
By the way, I think it is easy to miss out on an important trend if one tries to segment out 64-bit Windows users who want 64-bit “desktop” GIS away from 64-bit Windows users who are interested in 64-bit “server” GIS use. That is a very ESRI perspective but it is not a distinction so much made in the mainstream.
There are some ESRI folks who are being driven into 64-bit operation server side because that’s the only way open to prevent time-sharing in the old SDE/IMS/ArcGIS Server way from being a complete joke. Going 64-bit is the easiest way to get more jolt out of your server, so I can see why they are going that way. Makes sense.
But in the big mainstream world they usually don’t take very seriously a strategy that thinks a “stack” consists of only one small slice of 64-bitness out of very many things that are necessary. They understand perfectly well that in a rich organization/application going 64-bit means going 64-bit for all the links of the chain and doing so for the operating systems of the day. This “oh, yes, gosh, we have SDE going 64-bit now for XP and Vista” is pretty lame. They want it all 64-bits, no excuses, and for Server 2003 x64 and Server 2008 x64 as well.
47
cluster
// Jul 1, 2008 at 1:13 pm
Hi Scooter,
———————————————————–
Assuming Manifold is as good as Dimitri paints it to be, is there a reason not to do any of the things you mention either on the DB or on the client as he suggests? I can’t see any.
———————————————————–
That’s why I recommend you test/try all the options yourself, not just assume. If Postgresql + PostGIS works fine for your, that’s great.
But at least, in Oracle world, so far I cannot easily address the hard-coded SDO_ORDINATE _ARRAY size (You can check Oracle user’s forum for more information).
On middle-ware, we have to put it broader picture (not just ArcSDE). Do you think we should get rid of all kinds middle-ware built on top of RDBMS (maybe by having an almighty RDMBS, which has ALL the components everyone can ever imagine)?
I am not a fan of middle-ware (or whatever technology for its own sake). But middle-ware does have advantage in managebility. And for most part, I like competition, no matter Oracle, Microsoft, IBM, …, bring them all on!
I don’t want (don’t believe) there will be a “best” RDBMS/OS/middle-ware/programming language/…
48
James Fee
// Jul 1, 2008 at 1:25 pm
Yes a “rich” organization could do that. *rolleyes*
Just because my admin uses web mapping from our web services doesn’t mean she needs to run 64-bit. Please, there are reasons to go 64-bit, but this vertical stack needing to be all 64-bit sounds more like AMD/Intel talk than actual real world reality.
Yes my new Dell M4300 workstation laptop could easily run a 64-bit OS, but it won’t be in its lifetime, I can guarantee that.
49
Dimitri
// Jul 1, 2008 at 1:28 pm
JW,
Sure you do… you run across that need every time your 32-bit ArcStuff crashes.
There are three reasons most people who go to 64-bit operation are glad they did:
1. Reliability. 64-bit Windows running 64-bit applications tends to be a much more reliable environment than 32-bit Windows running 32-bit applications. Just having the benefit of many fewer crashes is worth it. There are three reasons why 64-bit applications are much more reliable. First, 64-bit Windows versions are more reliable than their 32-bit counterparts (a particularly noteworthy effect in the case of XP), so any crashes arising from Windows are greatly reduced. Second, many crashes in 32-bit programs come from programming errors in memory management that go away when 64-bit memory spaces remove the hard limits against which programmers bounce. It’s like setting off across a desert with a small gas tank that has a leak… you could run out of gas, but if you had a very much larger gas tank you could get across without ever realizing you had a leak. The third effect is that the line-by-line review required to re-write a 32-bit application to run native 64-bits tends to turn up a lot of errors that get fixed and so will no longer cause crashes.
2. Performance. 64-bit Windows running 64-bit applications is usually much faster than 32-bit Windows running 32-bit applications. I’ve yet to meet a user who once they get a taste of faster performance is willing to go back to slower response, longer pauses, more time required to do almost anything. There are many reasons why 64-bit applications in Windows run faster than 32-bit applications. This applies in even smaller data, but it can be dramatic with larger data. Redisplay of larger images or drawings, for example, is often twice as fast in 64-bits than it is in 32-bits, all other things being equal.
3. Capacity. 64-bit Windows running 64-bit applications can usually manage much larger data than can 32-bit applications. This is in addition to the implied ability of handling larger data caused by the first two points above. Many memory management bugs in 32-bits, for example, cause crashes so frequently when work is attempted with larger data that as a practical matter they can’t handle the larger data. Likewise, if it takes too long to do anything with larger data because of 32-bit performance there again you don’t have the capacity. But there is also a factor of absolute upper barriers to some types of data given 32-bit process space limitations in many applications that simply does not exist in 64-bits.
By the way, we’re not talking 8 GB as “big” data, because many effects come into play in as little as 500MB. 32-bit processes are limited to 2GB as a practical matter but given overhead and Windows stuff linked into the application you can easily top out at 1 GB. Given the queues and UnDo images and all the rest a modern GIS application will be maintaining internally and you could actually start hitting some limits at 500MB or even less.
That means that even though you have have plugged 8 GB of RAM into your motherboard (purchased for as little as $150 these days…), it could be that when you go to open that 500MB shapefile in a 32-bit application you could already be rubbing up against 32-bit limitations that start you down the sad road of paging to disk, process space limitations and all the rest that causes modern folks to curse the limits of 32-bit operations.
And, if you are working in a modern GIS environment you know how the rising use of images and image servers even in legacy settings is driving the content of projects well past the multi-hundred megabyte range. Getting over 500MB is no longer exotica, it is routine and even getting over 1 GB is no longer exotic but routine in many circles.
Getting away from all the problems of 32-bit limitations is trivially easy: you already have a 64-bit processor and the industry wants to sell you inexpensive RAM in dirt-cheap 64-bit multi-gigabyte modules, so for God’s sake use it. Run 64-bit Windows and use a 64-bit GIS.
50
Dimitri
// Jul 1, 2008 at 1:58 pm
James,
Intel and AMD don’t care because long ago they shifted to selling mostly 64-bit processors. Almost everything sold for professional desktop or server use shifted to 64-bit hardware a couple of years ago.
Most corporate / professional / enterprise desktops and servers are already at least Core 2 Duo, which are 64-bit processors, and the trend now is to buy Quad because you pay about the same. It seems to me kind of wasteful to buy a 64-bit processor and then throw away the 64-bit benefits it gives you by running it exclusively in 32-bits.
64-bit Windows doesn’t cost you anything more either, because as far as Microsoft is concerned that same product key works just fine whether you run 32-bit Windows or 64-bit Windows. They have no problem with that at all.
I’d say *especially* if cost is a factor you owe it to your organization to squeeze out every bit of value out of the equipment you’ve already paid for. Most likely you already have paid for the 64-bit hardware. Use it!
Not only that, but the greater reliability and performance of 64-bit applications will save you precious labor hours and yield more productivity. Just about everything runs faster on the desktop in 64-bits. Just as for servers the quickest way to get more jolt into your server is to run 64-bits and take advantage of lots of cheap RAM, that’s also the case on the desktop.
There’s no downside to this because you can still run your 32-bit applications in 64-bit Windows. It used to be the case that some third party hardware add-ons didn’t come with 64-bit drivers, but that sort of thing has become very rare in 2008. These days pretty much everything supports 64-bit Windows out of the box, and if some particular gadget does not it’s usually a small matter to find one of their competitors that does.
I will grant you that some vendors are slightly slower supporting Windows Server 2008 x64 with a wide array of desktop gadgets. And it’s true that some very new, experimental things like the latest CUDA items become first available for XP x64 and then only later for Vista x64. But even that tends to be a nit in most professional / corporate settings where the number of devices attached to computers tends to be the usual suspects (printers, scanners, USB gadgets, etc.) and not experimental hardware.
I also will grant you that some vendors have been slower to offer 64-bit Windows pre-installed even on their 64-bit hardware. That’s becoming rare in the case of desktop or server system vendors but it still happens in notebooks. Sony, for example, offers the odd spectacle of Vaio notebooks that are all 64-bit hardware with offerings of 4 GB RAM pre-installed but they will sell you only 32-bit Windows even though that won’t use all the 4 GB of RAM they offer to sell. If you want to buy Windows pre-installed and not hassle with an upgrade installation afterwards the Sony policy is a drag. No problem: buy a ThinkPad, which does offer 64-bit Windows pre-installed.
I guess the bottom line is that it’s not going to cost you anything, it will save you lots in terms of better reliability and performance and better use of hardware you’ve already paid for, and there is no downside because you can continue using your legacy apps along with newer 64-bit apps. Go for it!
51
RSF
// Jul 1, 2008 at 4:26 pm
James,
I’m amazed that you don’t see any advantages to ArcGIS desktop products going 64bit. Surely you’ve loaded some killer data sets in ArcMap and sat there staring at the screen waiting for display/redisplay. Granted, if all your data is coming from SDE, you won’t feel much pain, but for those who can only dream of using SDE, having the more robust computing resources brought to the table by a 64bit OS and 64bit ArcGIS would be a godsend!
Have you ever tried processing LiDAR data with ArcGIS and 3D Analyst (without a LiDAR specific extension) ? Try to create a TIN from over about 15 million points, open up task manager and watch as your RAM usage increases until ArcMap errors out and disappears.
Should you be lucky enough to get the LiDAR data to process without crashing ArcMap, you better be ready to go on a long walk to get a cup of coffee. It could take 10-20 minutes for the TIN to draw on screen. The GC industry jumped on the 64bit bandwagon years ago, and are reaping the benefits with special effects and 3D animations that simply were not possible/feasible in the 32bit world!
Granted, the average GIS user won’t take advantage of all the benefits that 64bits brings to the table, but the power user will immediately see the benefits! We should be pushing ESRI to go 64bit on the desktop, not telling them we don’t need it! Many of the performance bottlenecks that cause painfully slow redisplay of the screen with ArcMap would virtually go away if ArcGIS was a true multithreaded, 64bit application. Almost all of the crashes I experience with ArcGIS/ArcMap are related to lack of headroom in the 32bit environment.
(Disclaimer: I am an ESRI, Manifold, and 3D modeling/animation software user. I have seen first hand what 64bits does to improve both productivity and the user experience. I support the migration to 64bits 100%)
52
James Fee
// Jul 1, 2008 at 4:36 pm
@RSF: Of course you assume the only thing keeping ArcGIS from being a start is that it is 32-bit.
Look, I completely get the reasons why people want 64-bit GIS. It isn’t like ESRI isn’t working toward that goal, they are. Just the 9.x codebase won’t be there. They totally get that they need to be 64-bit, but the current release won’t be there (and neither 9.4). Everyone has told me 64-bit “will arrive” at 10.x so we’ll have to wait until next year to hear what that might be (or the UC?).
53
RSF
// Jul 1, 2008 at 5:17 pm
@James
Having experienced the improvements 64bit makes with 3D software (and Manifold), I’m a little impatient with ESRI. I lived through the “Workstation ArcInfo will never be ported to the PC” era, after the release of Win NT, and have always questioned their judgment on responding to dramatic changes in the computing industry. ESRI has been bit slow at responding to changes in a timely fashion. At least they are talking about 10.0 being 64bit, I just hope it really will be fully 64bit, and multithreaded as well. I also hope that 9.3 is a lot more stable, and a better performer than 9.2.
It will be interesting to see how Manifold and others fair in the GIS market over the next couple of years with ESRI behind the curve on the 64bit desktop.
54
M@
// Jul 1, 2008 at 6:13 pm
So back to the original point of the thread, we got our 9.3 dvds and EDN dvds today. No telling how long till we start installing them, however.
55
ManifoldUser
// Jul 2, 2008 at 2:05 am
@Dimitry: shut your mouth and bring your ass back to work. Manifold is a good tool among other but you don’t have to insult people to defense your point like that.
As a Manifold user, reading this thread make me shame. Growth up.
56
Connard
// Jul 2, 2008 at 6:25 am
ManifoldUser, I don’t see Dimitri insulting anyone in any of his posts in this thread. Your only post, however, contains an insult. In fact, I doubt that you are an actual user of Manifold.
57
ManifoldUser
// Jul 2, 2008 at 7:39 am
@Connard. This conversation is not constructive. Between the line, we all know Dimitry hate ESRI, and his voice here is quite a bit too loud.
I did use Manifold, and I think it’s a good of software (very flexible) that deserve a better public image. Jumping into a launching topic of a new GIS version just by saying how bad their product is compare to yours. In my standard this is an insulting to Manifold community.
58
Connard
// Jul 2, 2008 at 8:30 am
All due respect, I strongly disagree. Honestly.
Dimitri’s very first post started with:
“Look, if someone wants to praise ArcGIS 9.3 that’s OK, but let’s not have any nonsense about it being in any way a 64-bit thing. Praise it for other things but not that.”
…ended with:
“Despite all that, congratulations to ESRI for issuing 9.3 - not easy getting those big products out the door as every developer knows.”
…and contained no insulting rhetoric.
Many of the responses to that perfectly polite post did, however, contain insults.
I am all for having constructive discussion and I don’t see Dimitri faulting on that at all. The hate speech that you think originates at Dimitri’s posts in fact comes from other guys picking at him. Whatever their reasons, many of them don’t discuss the issues at hand and instead trivialize everything to “well, here’s Dimitri going again”.
There is no reason to be ashamed. Dimitri is perfectly in line and makes a lot of sense.
59
KoS
// Jul 2, 2008 at 8:33 am
@J…getting caught up on comments. I thought I made myself clear, I was talking about my part of the world. I can’t speak to what is or isn’t being used outside the state of Indiana or even outside of USDA as a whole.
You are correct, there are places where Manifold is in great use. Around here, it isn’t, unless people are too embarrassed to mention they use the program.
At the state wide GIS conference, the past 5 years, I’ve never ever heard the word Manifold used. Maybe they are all in the corner talking to themselves. Don’t know!
KoS
60
abridge
// Jul 2, 2008 at 9:24 am
ManifoldUser,
It was shrek, not dimitri who had the *stronger* tone this round, dimitri merely contributed his (albeit lengthy) opinion.
61
Buggy
// Jul 2, 2008 at 10:17 am
ESRI users are so touchy.
62
James Fee
// Jul 2, 2008 at 10:21 am
You are all so touchy.
It is GIS software, not College sports. No need to take anything personal as none of this really matters.
Now Arizona State whooping up on Georgia this fall, that is worth getting excited about.
63
KoS
// Jul 2, 2008 at 11:25 am
I call BS!!!!!!!
I think you excitment formula is a tad backwards.
This fall there will be plenty of water in Tempe, all those tears…
KoS
64
Buggy
// Jul 3, 2008 at 7:51 am
Thanks for proving my point James. Touchy, touchy.
65
KoS
// Jul 3, 2008 at 11:12 am
Buggy, you must be new around these parts. O’wait I guess I’m being soooo touchy. And I know your next thought, o’why, o’why, can’t we all get along! Argumentative doesn’t mean touchy. Hang out with lawyers, it may seem some are touchy, rather in fact, they love/enjoy to argue.
I guess you have missed the grief I’ve given James in the past. ASUers think they are hot shit coming off a great season with a good coach.
Sadly, this season a big Bulldawg is gonna to flatten them the first game of the season. They may not recover from the beating. The rest of the season they will have nightmares in red, white and black colors.
I take great joy continually reminding James of that fact! I want him to be completely prepared for the coming disaster.
Point not proven, don’t pass go or collect $200.
KoS
66
James Fee
// Jul 3, 2008 at 1:17 pm
Well we both did lose 2 of 3 to Fresno State in college baseball.
KoS, didn’t see Georgia above ASU in the directors cup this year. Sad that ASU finished third in the Pacific 10, but fourth overall.
You coming out for the game KoS?
67
bon
// Jul 3, 2008 at 2:18 pm
has any figgered out if the Data&Maps changed with the 9.3 release (other than a new package 7 copyrighting)? In particular the North America StreetMap…so convenient for geocoding - which by the way I find MUCH faster under 9.3. With a cute new little status screen too!
TIA! bon
68
KoS
// Jul 3, 2008 at 3:24 pm
Baseball???? Directors cup???? Bah!
We are talking football! Who cares about the other crap.