I’ve been reading up on some blogs over the past few weeks and a common theme is that because ESRI is “proprietary” they will fail in the long run and I just don’t see that happening. There is much confusion with how open some products are. For example many people say that because the Google Maps API is out in the open it will push ESRI off the map. I don’t see how Google Maps is any more open than ArcIMS or ArcWeb Services are. All 3 are commercial products and just because Google Maps is “free” (and to a lesser extent ArcWeb), that doesn’t mean that people will abandon paid for server products such as ArcIMS or ArcGIS Server any more than they abandoned Oracle or SQL Server for MySQL. In fact I’d say that Google Maps is more closed than ESRI’s products are because you can’t see what they are doing behind the scenes. Google is probably more secretive than ESRI is about their future plans and unlike ESRI they don’t even seem to listen to their customer (mostly because Google Maps customers aren’t paying a dime for tech support) and users are forced to create their own forums for support. Some would argue that this is a good thing, but I’d say being out on an limb like that could mean that Google could cut you off at any second and go in a different direction. ESRI hasn’t done this because their customers have direct access to the programmers at ESRI and they in turn listen to what people want making sure that the rug isn’t pulled from underneath them (take the continued “support” of ArcView 3.x as a perfect example).
Now what if we look at a true open product like UMN Mapserver? There is more support there for developers than there is for Google Maps. But does Mapserver create pressure on ESRI to conform to OGC standards? You bet it does because ESRI listens to their customers. I would also wager that the growth of Mapserver hasn’t really hit ESRI that hard as the total marketplace has grown so much over the past few years. ESRI is by no stretch of the imagination the first company one might associate with “standards-based open architecture”, but I believe their continued support for open formats shows that they do understand people want to connect to as many data sources as possible.
All the openness in the world won’t make any product successful, but listening to your customers will. The feeling that I’ve gotten from ESRI over the past year is that they have finally begun to realize that their road to continued success is supporting users like us. Don’t confuse the hype surrounding Google Maps/Earth with them being open and listening to their customers. There is no company that likes to hide behind their logo more than Google and they will do whatever it takes to not have to be open. There is a reason people are beginning to realize that Google is the next Microsoft (while Microsoft seems to have become the next IBM). Believe me, ESRI has a LONG WAY TO GO before they are as open as we’d all like them to be, but they do listen to their customers and that is a start.
You don’t feel like ESRI listens to you? Let me know in the comments.
3 responses so far ↓
1
Howard Butler
// Aug 26, 2005 at 8:04 pm
I agree that MapServer is probably complimenentary rather than an either/or proposition for most organizations. Those that typically use ESRI products, enjoy them, and have had success (egoboo from bosses and clients and the like) will most likely continue to do so.
Organizations without the resources to attempt developing completely within the ESRI software stack can dip their toes in the water with the Open Source software stack with relatively low impedence. Without the Open Source alternative, they might not even bother at all with a mapping project. This is a boon to small-time consultancies that can price themselves without respect to licensing fees, keep the cost of the entire project under budget, provide value to customers, and still make it worth their while. In my opinion, there is room in the marketplace for overlooked jobs that a one- or two- man group with some expertise can knock out if the consultancy doesn’t *also* have to eat the cost of licensing (clients in my experience tend to think total project cost, not merely $$$/hour for customization and development). Specialty stuff like environmental compliance modeling, water rights benefit assessment, and specific remote sensing problems are reasonably scoped in size to fit this classification.
That said, there are still some things that MapServer does that (I hope) puts some competitive pressure on ArcIMS et. al that benefits the community as a whole. I implemented support for ArcSDE versioned queries in MapServer because at the time (it might be able to now), ArcIMS didn’t have a direct way to support doing so. Surely not everyone needs this feature, but I did, and MapServer (and the ESRI ArcSDE C API) gave me a way to implement it. The ability to cascade WMS and WFS is also something that I would really like ArcIMS to be able to do, and in time, hopefully it will be implemented (I assume ArcServer already supports this). Duplicating data wastes everyone’s time, and I really like to let the folks who are “in charge” of maintaining that data continue to do so, provide the data in a way that supports on-demand query, and free me from burning my resources with storage, maintenance, management, and updates of stuff I really don’t have any control over and is a sunk cost.
The ArcWeb Services vision seems to get me there, but as a small-time developer, AWS makes me a bit uncomfortable because if some govt. agency mandates “you can only use X to access data Y,” I may be in a position where I have to re-implement things again. Of course, you could also argue that the OGC is as much if not more of a moving target and as a developer you may still be in the same boat. Pick your poison.
As a recent thread on the geowankers list discussed, Google Maps has sucked up a lot of the developer “mindshare” with regards to internet mapping. I don’t know that it is necessarily true that it has sucked up a lot of mindshare from typical GIS web developers (which is probably a relatively small group in relation to the potential size of the new-to-GIS-and-mapping Google Maps developers), other than as a passing curiosity. At least, it hasn’t for me. The thing that makes it so disruptive is that it only requires a developer’s key to work with and it covers X% of the web GIS problem domain, where X% is the list of things that most web maps need to do — plot points, pan, zoom, geocode, and label. However, just as LAMP (Linux, Apache, MySQL, Perl/Python/PHP) didn’t make the WebObjects, Cold Fusions, and JSP/ASPs of the world extinct, neither will Google Maps/Earth and/or MapServer/GeoServer/PostGIS/GRASS make the ESRI or AutoCAD software stacks go away.
The discouraging thing to me about Google Maps is that it means another standard (KML), another development stack (AJAX), and another way of doing things. You can also argue that this is a good thing. Either way, you can’t ignore it, however. If you are a data provider, you’d better make sure your software emits KML so that the Google Earth users can drop your data in their interface. Otherwise, a significant (if not a majority) number of potential users might not know your data even exists.
Oh, one more thing. I’m proud to admit that I still use ArcView 3.x almost daily
Viva la’ Avenue!
2
Matt Perry
// Aug 27, 2005 at 1:36 am
>> All the openness in the world won’t make any product successful, but listening to your customers will.
I totally agree! Proprietary, open-source, web services, whatever kind of solution you choose, the key is that the developers must stay in touch with those that USE the software on a day-to-day basis. In a sense it all comes down to building a community around a particular technology… the software packages without strong ties to their user base will miss out when new technologies arise.
While I love open standards in theory, Google Maps shows that you can get RESULTS while completely ignoring the status quo. In the end that’s why we’re in the GIS game.. to get results.
But I never want to see the day when I am FORCED to use a particular technology in order to access my data. Public data should always be distributed in a format that can be used in any GIS software (either directy or through the use of conversion software).
3
matt
// Aug 29, 2005 at 9:47 am
While it is true that ESRI has open APIs, the thing that is getting so much developer mindshare oriented towards Google’s technology is that they have put the data out there as service. Code that people write to hit the Google APIs can be run against free data. Google’s free data is better than the data most places can afford to purchase.
ESRI is fast becoming the software used to create the data, not the software to use it. They need to get behind a big, free data service now.
Leave a Comment