James Fee GIS Blog

39 responses so far ↓

• 1 Cellulose // Jun 18, 2007 at 11:09 pm

  I think the Planet Geospatial got most of it right, but is a little off on a few things. As a programmer, none of that news was useful to me either. Quite frankly, I’m just here this year to corner some SDE developers and collect business cards so I can bypass the front-line support.

  Up front, Jack admitted they weren’t adding much to the product suite for 9.3. And the demo’s reflected that–a few usability tweaks here and there. And, of course, a reminder of a few products they haven’t showcased in a couple years in case you had a few grand to spare.

  New features? I think I saw about six new features. Everything else was a reminder of already-done features. I was able to very several of the features already existed in 9.2 (which were not called “new” BTW).

  Sebastian keeps perpetuating this notion that ArcSDE won’t support native PostGIS types (i.e. “shooting the bird to the entire PostgreSQL spatial community by not adopting their spatial type”), even though nothing has been announced to support his claim. The ESRI reps and devs I’ve talked to pre-conference have all stated that native PostGIS will be supported.

  Some of the “spin” that Sebastian put on the ESRI comments and demos I think are a bit one sided or simply don’t make sense. Amateurs can go #$() themselves? I’d definately call ESRI pretentious and arrogant, but I don’t think they sent that message.

  Tell me–has anyone ever PRODUCED data using Google or Microsoft? And if they have, can you point me to a tutorial on creating data?

  When we talked to Google a few months ago about their fusion server, they flat out told us to go somewhere else–GE can’t “produce” data. So amateur or pro, I don’t see how “publishing” data to not-really-free-consumer services is telling a huge segment of their customer base to go #()@( themselves. Maybe I’m just slow… or maybe I’ve just drunk too much koolaide and now brain-damage.

• 2 James Fee // Jun 19, 2007 at 6:39 am

  Don’t you think we got this talk last year with 9.2?

  I’m getting tired of the same song and dance. Either fix the bugs or don’t. ArcGIS is mission critical to my workflow and its not up to the task.

  Plus he gets points for actually having an opinion. Just blogging bullet points is a waste of time.

• 3 KoS // Jun 19, 2007 at 7:03 am

  I couldn’t agree more……”Either fix the bugs or don’t”.

  I get tired of the newer version to fix previous bugs template ESRI tends to use.

  KoS

• 4 David Davis // Jun 19, 2007 at 7:14 am

  I too am growing tired of the same spin from the UC. Sitting there I could only think that I’ve seen this exact speech before many times. The smiles are nice, but how about fixing the damn product.

• 5 Ian // Jun 19, 2007 at 7:33 am

  James: “ArcGIS is mission critical to my workflow and its not up to the task.”

  Seriously. We’ve been working with ArcGIS Server now since Beta 2, and it’s just inexcusable the number of stability problems it has. ESRI Marketing sold it very hard as “Enterprise GIS” and it’s simply nowhere close, nor do I expect it to be until maybe 9.3 or 9.4.

  The jokes on us for using it I guess. At least Jack was nice enough to apologize at the Plenary, or have someone apologize on his behalf. I’m sure he doesn’t feel badly enough though to refund my client the hundreds of thousands of dollars they’ve spent building a system on top of alpha-quality software however.

• 6 James Fee // Jun 19, 2007 at 7:38 am

  Ian, he apologized last year.

  Seems like that is going to be an integral part of the plenary going forward.

• 7 Jud Aster // Jun 19, 2007 at 7:49 am

  The smiles are nice.

  The international ESRI UC is such a fantastic event. Marketing and religion all together, people is happy and everything is perfect, don’t be so mad… The GIS world is getting too wide, and nobody can do it perfect, not even ESRI. But they are too expensive, yes.

  Never tried Geomedia or MapInfo ? ha ha

• 8 dylan // Jun 19, 2007 at 9:03 am

  “Either fix the bugs or don’t. ArcGIS is mission critical to my workflow and its not up to the task.”

  Why is the Arc series the forerunner in commercial GIS, then, if ESRI doesn’t fix bugs? I’m a GIS young’n; can anyone answer that for me? Why is our industry relying on faulty software? I’ve got ideas, but am curious what some more experienced folks are thinking.

  dylan

• 9 Cellulose // Jun 19, 2007 at 10:10 am

  In previous years, the “bug fixes” were one of the 100+ new features. Honestly, how much time can you spend on bug fixes if you have release with 100+ new features?

  Will it be any different with 9.3? Probably not.

  Shaved ice is less filling. Do I get a points for having an opinion too?

• 10 James Fee // Jun 19, 2007 at 10:21 am

  Question for you Cellulose:

  At what point do you just say lets fix things and forget about adding the kitchen sink?

• 11 Cam W. // Jun 19, 2007 at 10:26 am

  @Dylan: Why is Microsoft the forerunner in operating systems when they don’t fix the bugs in their system(s)? The answer is it’s the software that everyone knows how to use and is most likely the software they were trained on.

  The similarities between MS and ESRI go further as well. Rather than concentrating on fixing bugs and stability both companies are pushed to rush new features out the door in order to keep up with their competitors (who often give away or sell their products for much less money). Both ESRI and MS also suffer from the fact that both of their applications are used by a very wide audience (which leads to huge lists of ‘feature requests’) and their code bases’ are ridiculously complex. It’s not like ESRI can’t afford some really good programmers. They are just hamstrung with the mess that they have got to work with. I’ve been told (I’ve got no hard evidence to back this up) that ArcGIS / Arc Objects is the largest COM based application in the world. It’s supposed to be orders of magnitude bigger than what MS Office was (err…. is? Is MS Office still a COM product or has it gone to .NET??).

  I agree with Cellulose that Sebastian’s comments are a little mis-leading in regards to PostgreSQL support and a little over the top in general. The main issue that Sebastian raises’ is a good one though. The speech for 9.1 and 9.2 appear to be almost the same as 9.3…. “Sorry about the bugs. The next version will fix those problems, in mean time check out the cool spinning globe!”

• 12 Ian // Jun 19, 2007 at 10:36 am

  Dylan, to take a stab at answering your question, my view is this:

  ESRI has been the historical leader in analytic GIS. They essentially have had a soft monopoly on the field, and as the field has exponentially grown with advances in technology, ESRI has been well-positioned to maintain that lead.

  Organizations that have already invested large sums in SDE, ArcMap, etc etc, are ill-advised to switch horses.

  However, their quality control is terrible. They are foremost a “GIS” company, and a software development vendor second (maybe last?). As they continue to introduce new functionality and more complex architectures, their lack of software development practices becomes ever-more glaringly obvious. (see the disastrous TOC-patch which had to be released, then withdrawn, then re-released for Service Pack 2).

  ESRI’s “quirkiness” may have been acceptable when their client-base consisted of large governmental entities and municipalities, but as GIS is adopted by the private sector, the standards for performance are much more strict. ESRI may find they were better off sticking with academia and federal agencies.

• 13 Gretch // Jun 19, 2007 at 11:16 am

  I’m kinda tired of the same old thing. This dog and pony show has got to stop and ESRI needs to start responding to the millions of users.

• 14 Lefty // Jun 19, 2007 at 11:19 am

  You know Gretch, ESRI doesn’t give a rat’s ass about you. They only care about the top 20 companies that use their product. The fact that they are nice enough to let you use their product is good enough.

  That is the big myth about ESRI. They are no longer about users, but about consulting. Most of these “features” are for them, not you.

• 15 artlembo // Jun 19, 2007 at 11:23 am

  ESRI Marketing sold it very hard as “Enterprise GIS” and it’s simply nowhere close, nor do I expect it to be until maybe 9.3 or 9.4.
  

  @Ian: what do want to do to be able to call it an Enterprise GIS. That is, what is it not doing properly, and what do you think it needs to do?

  A bulleted list would be great. I’d like to give some thoughts to it, if you are able to thoughtfully list out some main ideas.

  thanks!

• 16 G Money // Jun 19, 2007 at 11:39 am

  Rather than whining on a blog site, does any of you ever pick up the phone and call ESRI tech support and report the bugs? If these are critical bugs in your workflow, do you ever ask to have them escalated? Here’s an example of what I’m talking about:

  “I’m kinda tired of the same old thing. This dog and pony show has got to stop and ESRI needs to start responding to the millions of users.”

  Really? Millions of users? Do you think if millions of users were actually reporting bugs that Jack would just sit back and let developers go off willy-nilly and pack in a whole bunch of buggy new features?

  Yes, ESRI does have a responsibilty to fix bugs. Bitching about how buggy the software is on a blog isn’t helping them to get the job done.

• 17 James Fee // Jun 19, 2007 at 12:45 pm

  What would help them G Money?

• 18 dylan // Jun 19, 2007 at 1:01 pm

  I’ve always assumed there were primarily two things that catapulted ESRI to the top: user base from ArcWorkstation days and no strong competitor sprung up in time to combat their infiltration into academia, which (just like with Microsoft) catapults standards if they get in early enough. But, like I said, I wasn’t around then.

  Ian-
  As far as why there are so many bugs now, that makes sense. I’ve seen ESRI as primarily a developer and secondarily as a GIS company. Hadn’t considered the other way around. Also makes sense that such a wide spread of industries would be very difficult to perfect their software.

  G$,

  “do you ever ask to have them escalated? “

  Yes, but if I emailed and called and asked and requested every time I came across a bug, I’d get even less work done than I do bitching on a blog.

  I don’t think millions are reporting bugs. That would help, a lot; I agree. I do think new features sells more products, though. I suspect most people assume a company as ingrained as ESRI is has base software that already works. What are the salespeople going to say? “Nope, nothing new or fancy this year, but we did manage to make this version actually work smoothly!” Flash gets the money in the bank, and like someone else mentioned, once that investment has been made it’s tough to start over with another environment- even once the bugs start rolling in. Throwing good money after bad money is easier to explain than admitting the first investment was bad money.

  dylan

• 19 Ian // Jun 19, 2007 at 1:10 pm

  @artlembo: Let me clarify. ArcGIS Server is extremely impressive from a functionality standpoint. I don’t mean it’s missing bells and whistles; the system we’ve built using it is extremely impressive to our users.

  When I say that it is nowhere close to being Enterprise GIS though, what I mean is the maturity of the product as a whole is lacking.

  Specifics:
  - Lacks robust API documentation
  - Very rudimentary, sometimes non-existent, error handling
  - Exhibits very little, if any, fault tolerance
  - SOCs appear to have memory leaks
  - Map Services frequently crash for no apparent reason
  - etc (see EDN AGS forums for many, many others issues)

  These are the sorts of things that distinguish a “feature-rich” software program from a mature, enterprise-capable software product.

  Fred Brooks describes this much better than me in Mythical Man Month. Check out the program vs. product vs. programming system distinction here: http://www.cs.usfca.edu/~parrt/course/601/lectures/man.month.html

  IMO, ArcGIS Server 9.2 is an extremely impressive “program” which is being sold by ESRI as a “programming system product”, and they are distinctly different entities.

• 20 Mark // Jun 19, 2007 at 2:10 pm

  Hmmm…Lefty brought up a good point. I wonder if there is vailidity to that statement. As a previous GIS Product Manager for a sofware company, I can attest to the fact that there is a lot of dough to be made via consulting…It made me mad because I would never get the cash I needed to fix bugs.

• 21 Cellulose // Jun 19, 2007 at 2:38 pm

  James–considering they now have 6 kitchen sinks in every bathroom and hallway, they should have stopped a long time ago to fix things.

  I am not disagreeing about the QA problems with ESRI… I just think there’s a tiny bit of hope (not that I would bet money on it or anything) for 9.3 since there’s far fewer features being added compared to 8.3, 9, 9.1, and 9.2.

• 22 Cellulose // Jun 19, 2007 at 2:48 pm

  Ian–
  Government and muni accepts bad performance, but private does not? That’s completely contrary to my experience…

  Ever used Oracle? PeopleSoft? APS? QuickBooks? Symantec Antivirus? Eudora? All tools embraced by the private sector and they’re as bad as ESRI in terms of quality, usability, and performance.

  Private sector will accept and embrace buggy, slow, and unusable products all the time. And the private sector will let themselves be extorted out of millions in consulting fees instead of using a viable alternative. Ask any business who made the mistake of adopting Oracle and can’t afford a full-time, quality DBA.

• 23 James Fee // Jun 19, 2007 at 2:51 pm

  But Cellulose: 9.2 was supposed to be that release. It turned into almost a total rewrite. Who says 9.3 won’t turn out that way?

• 24 Ian // Jun 19, 2007 at 2:53 pm

  Cellulose-

  I would respectfully disagree. Having done consulting and application development for many years both with Fortune 500s and with the Federal Government, my experience has been that the Federal Sector at least is extraordinarily tolerant of poor performance across the board.

  As far as middleware is concerned, I’ve built systems on top of Oracle, SQL Server, Informix, HP OpenView, etc. None to date have been as routinely unstable as ESRI.

  YMMV.

• 25 NoLove // Jun 19, 2007 at 3:32 pm

  By the time ESRI fixes all of the bugs in 10 (not 9.3) we will be storing and writing all of our spatial data to SQL Server Yukon via Virtual Earth or maybe some KML/WFS feature with Google Earth/WorldWind. Also until ESRI does something about the slow render times at the core level…it will all just be lipstick on a pig.

  It seems that the simple things need to be fixed rather then focus on new features.

• 26 G Money // Jun 19, 2007 at 7:53 pm

  James,

  “What would help them G Money?”

  Report the bug. Provide specifics, reproducible cases, with sample data, if possible. Explain the overall use case and workflow. Let ESRI tech support know how critical the problem is to your organization. Include your account rep. in the exchange, and let him or her know how critical the problem is to your organization. Talk to ESRI developers at the UC and hold their feet to the fire about bugs (provide bug numbers and/or incident numbers) that have been reported and not fixed, and let them know how critical it is to your organization.

  There are hundreds of feature requests, thousands of tech support incidents / bug reports, and tens of thousands of users. In my experience, ESRI folks genuinely want to do the right thing for their users, but how do they determine what really needs to be done?

  For ESRI, it’s like trying to filter out a signal from a whole mess of background noise. General complaints about quality or performance or pet features, through channels that ESRI folks may or may not monitor, just add to the noise. The challenge for the user community is to communicate the signal to ESRI as specifically, succinctly, and directly as possible.

• 27 Cellulose // Jun 19, 2007 at 11:04 pm

  Ian–

  I think our sampling sets have zero overlap, hence the wildy different viewpoints. I work at a federal R&D lab and most of the work we do for other agencies have performance requirements that far exceed commercial and private-sector requirements…

  So I think you are right about many federal agencies/projects. I just have the good fortune of avoiding those projects and working on the really challenging and interesting projects.

  So where does that leave us with ESRI? We use it where economical to free up resources to work on the really hard problems.

• 28 Cellulose // Jun 19, 2007 at 11:34 pm

  James–

  I disagree with your assessment of the 9.2 release. Almost from day one, the 9.2 release was presented as a feature-release. QA was feature 231,295 of the one million new features for 9.2… How could that be The QA Release?

  The marketing of 9.3 has been trying to set a different expectation than with 9.2. If 9.3 is a major rewrite, they’ve done a terrible job setting up an aliby for when it fails to meet expectations.

  At least 8.3 through 9.2 included huge “new features” lists to distract us from The QA Release.

  Either way, I’ll satisfied as long as they fix the bugs I’ve submitted.

• 29 James Fee // Jun 20, 2007 at 6:50 am

  Cellulose:

  How quickly we forget:

  ArcGIS 9.2 was mentioned for a early 2006 release. Clint Brown was then featured in a small video explaining that the 9.2 release would be consumer driven, more usable, higher quality with more documentation.

  By the 2006 UC, 9.2 ceased to be this type of release, but originally it was to be “high quality”.

• 30 KoS // Jun 20, 2007 at 6:56 am

  Don’t forget, almost every Federal agency has it’s own culture and business practices. Some are better than others at doing the projects right. Even within a single Dept, their sub-agencies generall have different cultures and practices. Which I find somewhat weird, but understandabe once one looks under the hood.

  Can’t paint using a broad brush when talking about the Feds.

  I’m still amazed the system hasn’t crashed and burned by now.

  KoS

• 31 Cellulose // Jun 20, 2007 at 7:39 am

  James–

  No, I haven’t forgotten–your summary pretty much matches my notes from the same plenary… My point still stands–quality was just one of several aspects of the 9.2 release. By your own UC2005 summary, it’s clear that 9.2 was a feature release, not a quality-release. The word “quality” thrown in a couple times throughout the day doesn’t constitute a commitment to quality.

  The *only* thing that makes 9.3 different is they’ve promised a whole lot less in terms of features–what will they give us to appease our anger if they fail to live up the QA promises?

• 32 James Fee // Jun 20, 2007 at 8:50 am

  The promise of ArcGIS 10.x?

• 33 Cam W. // Jun 20, 2007 at 8:55 am

  I’m not at the UC this year but in response to Cellulose does anyone have a list of ‘new features’ that are coming in 9.3? I know sometimes ESRI likes to confuse new features with bug fixes but this should be pretty simple. I honestly can only think of a couple of new geoprocessing tools that are coming in 9.3. I’m not aware of any major new features.

• 34 Cellulose // Jun 20, 2007 at 5:57 pm

  LOL.

  Touche James. You win this one.

• 35 top // Jun 20, 2007 at 9:19 pm

  Somewhat tangential to the original topic:

  @Cam:

  I’ve been told (I’ve got no hard evidence to back this up) that ArcGIS / Arc Objects is the largest COM based application in the world. It’s supposed to be orders of magnitude bigger than what MS Office was …

  Do you mean the number of lines of code? If so, this is extremely unlikely. Office is huge, tens upon tens of millions of lines. I doubt the codebase for all ESRI products combined is much larger than five million lines.

  … (err…. is? Is MS Office still a COM product or has it gone to .NET??).

  It has not gone .NET in that the vast majority of the code is unmanaged. This is not going to change anytime soon and quite possibly will never change at all.

• 36 James Fee // Jun 20, 2007 at 9:42 pm

  @top: ESRI used to say ArcGIS was the largest COM based object model.

  They’ve toned it down to “one of the largest COM-based commercial componentware toolkits available anywhere”.

  Either way it is a beast and has become too large for most folks (which is why ArcGIS Engine is now available).

• 37 Canuck // Jun 23, 2007 at 7:08 pm

  Well, after all the discussion, what can be done about it? Is there an alternative? Is there anyone else with the spatial analysis and other tools that can be implemented widely, other than ESRI. I mean, GRASS is great and can do a lot, but it certainly will not suffice for an enterprise implementation the way ESRI software does.

• 38 top // Jun 24, 2007 at 11:37 pm

  Is there anyone else with the spatial analysis and other tools that can be implemented widely, other than ESRI.

  Is that a softball? Of course, there is an alternative. See this blog of James:

  http://www.spatiallyadjusted.com/2007/06/14/manifold-returns/

• 39 How do I do that in ArcGIS/Manifold: illustrating classic GIS tasks « Free James From ESRI // Jun 25, 2007 at 7:56 am

  [...] extensions that are standard with Manifold Professional. Anyone, like some of the posters over at James’ blog, looking for a real alternative to ESRI products may be convinced by this document that Manifold is [...]

Leave a Comment