Fracking Legacy GIS

January 31st, 2008 · 29 Comments · GIS

What they heck is legacy anyway? Look if it ain’t running on VMS, it isn’t legacy in my opinion. Most people use “legacy GIS” when referring to ESRI, I get that because it is just another buzz word. Rather than tell me why I should use their product they just focus on the problems with ESRI ArcGIS, Oracle Spatial, Intergraph and not why their product is better.

Tags:

29 responses so far ↓

1 atanas entchev // Jan 31, 2008 at 4:15 pm

Most GIS people I have talked to use “legacy system” when referring to ArcInfo workstation, a.k.a. ArcINFO, ARC/INFO, etc. AML is also a good example of “legacy.”
2 James Fee // Jan 31, 2008 at 4:20 pm

ArcStorm would be legacy, not ArcInfo workstation. Using legacy to refer to ArcInfo Workstation is succumbing to buzzwords.

Legacy refers to out-of-date software still in use. ArcInfo Workstation is not out-of-date. ArcLibrarian is legacy.
3 bugaboo // Jan 31, 2008 at 4:28 pm

I think “legacy” is poor terminology to describe any ESRI products. “Traditional” would be a more logical term to use for describing the ARC product line.
4 Lefty // Jan 31, 2008 at 4:43 pm

I think folks refer to ESRI as Legacy in marketing terms. I use a legacy database every day, FoxPro 2.6!
5 atanas entchev // Jan 31, 2008 at 4:50 pm

According to Wikipedia, “A legacy system is an old computer system or application program that continues to be used because the user (typically an organisation) does not want to replace or redesign it.”

ArcInfo workstation (NOT the ArcInfo flavor of ArcGIS desktop) and the coverage data model are legacy, in my book, and according to the Wikipedia definition. The user (the federal government) does not want or cannot port/migrate their AMLs and coverages to a modern environment. They are using a legacy system.
6 James Fee // Jan 31, 2008 at 4:51 pm

But AMLs still run in 9.2 without modification. Soon one could say Avenue is legacy once ESRI stops supporting ArcView 3.3.
7 atanas entchev // Jan 31, 2008 at 5:04 pm

But Avenue scripts will continue to run after ESRI stops ArcView GIS 3.3 support. Is vendor support (or lack thereof) what defines a “legacy” system? Methinks not.
8 James Fee // Jan 31, 2008 at 5:07 pm

If their reasons for staying on ArcView match what you said above it would be. Just because AML is “old” doesn’t mean it is legacy. FORTRAN is older than any of us here, but it is still used for GIS analysis. Is that legacy GIS?
9 atanas entchev // Jan 31, 2008 at 5:16 pm

I am reminded of the attempts of the US Supreme Court to define “obscene”, or hard-core pornography. In 1964, Justice Potter Stewart tried to explain “hard-core” pornography, or what is obscene, by saying, “I shall not today attempt further to define the kinds of material I understand to be embraced… but I know it when I see it…”

To take a page out of Justice Stewart’s book, I know legacy when I see it. Carburetor is legacy; fuel injection is not. Black-and-white TV is legacy; color TV is not. ArcInfo workstation is legacy; AGX is not.
10 Brian Timoney // Jan 31, 2008 at 5:42 pm

Out of the mouth of a salesperson, “legacy” means anything that was in use the day before their particular product launched.

A more useful context that I hear it cropping up in is when companies merge and unintegrated “legacy” systems from one of the companies (the smaller one, typically) that the new entity doesn’t plan to support, but that people actually use to get stuff done day-to-day.

A lot of the time, though, it’s merely an acceptable way to browbeat someone into thinking they’re way behind tech curve….

BT
11 Dave Smith // Jan 31, 2008 at 7:34 pm

To me, “legacy” typically implies a monolithic technology stack, which doesn’t play nice with other technologies, or on a broader level, something that’s just not well-integrated or which is stovepiped - however that in and of itself isn’t necessarily a bad thing. For example, I still know of a few MapObjects IMS sites which still perform faithfully without crashing or maintenance, and which meet their intended needs, just as I know of shiney new AGS or other current-technology sites which were already obsolete before they even went into production.

It’s one of those words which strives to give a negative connotation, but not always justifiably so.
12 j // Jan 31, 2008 at 9:35 pm

To me, I see it as being quite simple and that we are trying to define the wrong term.

Invariably after some “thing” is created, a “thing” superior to the initial one is eventually created at a later date. Who this is done by does not factor into it.

This is because that first creation had its purpose and limitations based on the working environment at the time. The assumption is that these conditions and environments improve over time, thus giving rise to something superior. The intent and expectation is that one would (obviously) want to use the new superior creation.

That being said - use of the now inferior creation *should* be phased out. But while it is still in use, it would be considered outdated, or legacy:-) Not saying it can’t be used, or shouldn’t be used, just that its not the best tool for the job anymore, and often its continued use can hinder advancement.

So, the real question is not to define “what’s legacy”. The real question is to define “superior”;-)

My 2 cents.
13 Andrew de klerk // Jan 31, 2008 at 11:08 pm

The Manifold users love “legacy” products…..A search of their forum reveals 98 odd posts with the word “Legacy” somewhere in the text - “Legacy Shapefile format” etc etc.

I think you hit the nail on the head…yes there is a lot of marketing that goes on by saying other vendors have legacy products. One of my biggest gripes about the manifold community is their constant disdain for ESRI and their products……heck and I am a manifold users as well!
14 James // Feb 1, 2008 at 3:09 am

To a Manifold user…..legacy is ESRI!

Whilst it may be true that in terms of a GIS user base, ESRI may be larger, but it seems that ESRI users spend their life discussing how rubbish the products are, how much they crash, how expensive they are etc etc. Very rarely do I come across a satified user. I agree with the comment made previously that legacy means monolithic software stack that doesn’t work properly, is expensive and does not play nicely with others.

Manifold on the other hand is cheap, fast, reliable (I have never had Manifold crash on me in 3 years of use), comprehensive etc etc.

A Manifold user is a happy user and we never tire with what we can achieve with the product.
15 ewell naimbo // Feb 1, 2008 at 5:28 am

legacy or not, arcview 3.3 still opens, loads and pan shapefiles faster than any present gis system!
16 James // Feb 1, 2008 at 5:32 am

I should think that’s another good benchmark for legacy….anything that relies on shapefiles to get things in and out of a GIS system.
17 mr. legacy // Feb 1, 2008 at 6:44 am

legacy is not always a bad thing..

Joe Montana would be considered a legacy. The billing system used by a company might be a legacy system - written in COBOL, but is 100% reliable - everyone gets their checks after two weeks.

The upstart computer firm offering a new billing system to an accounting firm usually has major problems initially - that is why people are so reluctant to move off the “legacy” system.

So, newer systems that can talk with a legacy system are a good thing.
18 atanas entchev // Feb 1, 2008 at 8:23 am

I want to throw my support behind mr. legacy above.

Yesterday we started a discussion here about WHAT IS legacy. I wake up this morning, and I see that the discussion has veered into HOW BAD exactly legacy is. And I don’t think it’s bad at all.

Let me offer a quote from Walter Kirn’s article about (against) multitasking, from an episode in which he wants to buy a cell phone without a camera (I recommend reading the entire (long) article; you will thank me later):

““I’ll take the fat little free one,” I told the salesman.

“The thing’s inert. It does nothing. It’s a pet rock.”

I informed him that I was old enough to have actually owned a pet rock once and that I missed it. “

Here is the full article:

http://www.theatlantic.com/doc/200711/multitasking
19 Bill // Feb 1, 2008 at 9:06 am

IMHO, “legacy” is a completely subjective term. It means whatever the person using it thinks it means (or the salesman Brian mentioned above).

I agree that legacy systems are not necessarily bad. It’s interesting that COBOL was mentioned. A friend of mine works for a regulatory agency that still uses millions of lines of COBOL for one simple reason: it still works correctly.

One of the hardest arguments to make in any shop is making the case to shift to a new platform. People develop a sense of trust a familiarity with software (warts and all) and it’s hard to shake that.

I think that’s what ESRI has been experiencing with ArcGIS Server. Despite all of the marketing, people realize it’s a completely new platform that requires significant migration.
20 George Silva // Feb 1, 2008 at 10:11 am

“Ah ESRI is bad. ESRI crashes, ESRI sucks”

Why so many people use it?

Best documentation i´ve ever seem.
Good products
Expensive, true.
21 J Wallis // Feb 1, 2008 at 10:24 am

yeah, you got to hand to ESRI, they probably have some of the best technical writers on the planet.
22 Dave Smith // Feb 1, 2008 at 10:48 am

@George Silva: Why do so many people use Micro$oft products, knowing there are flaws and security holes? It’s not always just about functionality, documentation, and so on…
23 George Silva // Feb 1, 2008 at 10:52 am

@Dave

I dont know. If i did i would charge a lot of money to answer it :p

“peverting” the discussion, ArcGis is some miles ahead of any other product (even with all the problems), dont you agree?

I never tried Manifold, but it seems good to. At the moment i have a licensed AGIS, so i´ll stick to it.
24 Doug // Feb 1, 2008 at 11:49 am

Legacy:

That which is no longer supported by the original vendor, but which is still regularly used.

Windows 2k would be a good example of a “legacy” operating system. (this isn’t a dig against win2k which was a good OS).

My 2 cents.
25 Gady // Feb 1, 2008 at 11:58 am

“Rather than tell me why I should use their product they just focus on the problems with ESRI ArcGIS, Oracle Spatial, Intergraph and not why their product is better.”

Reminds me of a presidential election year!
26 Andrew de klerk // Feb 1, 2008 at 12:48 pm

Is WordPerfect a Legacy System? I came across someone the other day still using it. Remember the old blue and white of WP 5.1!
27 atanas entchev // Feb 1, 2008 at 1:04 pm

My 30-year-old mechanical Seiko watch is a legacy system; does the job better than most new quartz pieces of … hmmm… plastic and silicon.
28 FroodCNB // Feb 2, 2008 at 7:52 am

Trying to define the word “legacy” is almost like trying to define the words “old”, “big” or “useable”. It’s a relative term. It imples older but useful, something to keep using while also using newer things. It is not a negative term unless it’s spun that way. In contrast, “Version 1.0″ is a negative term.
29 tg // Feb 7, 2008 at 7:12 am

There only two “legacy gis” applications in my world.

Microstation with MGE and PC Arc/Info - thats about as legacy as GIS can get and still have been post VMS.