I’m following various discussions about layered web applications–that is, software that “layers over” other web sites or software. That includes “mash-ups,” some API-based applications, and lots of other “Web2.0ish” things.
Lots of these ideas and applications are wonderful. Once in a while, I do have mild skeptical thoughts about two aspects of them–particularly if and as such applications are suggested as replacements for more, shall we say, traditional applications rather than as extensions or complements.
I’ll just mention one in passing, since it should be obvious to anyone who’s been through the dotcom bust: A layered application ceases to work if the underlying operation goes away. But you all knew that, right? Whenever any private business says “forever,” be a little cautious: “Forever” can mean “at least through the end of this fiscal year.” (The same is true for nonprofits and government entities, of course.)
My other mild concern is highlighted in the posting title. To wit:
What’s the ‘business’ relationship between this layer and the underlying operation?
I think there are three general answers:
- Best case: Symbiotic. The layered application clearly benefits from the underlying operation, but in a manner that also benefits the underlying operation.
- Most common case: Mildly parasitic. The layered application uses some of the underlying operation’s resources without providing any benefit to the underlying operation–but the amount of resources used is relatively small, and the layered application doesn’t weaken the underlying operation except to the extent that load becomes a problem. (Not that load can’t become a problem; very few underlying operations have the apparent robustness of, say, GoohooMszon.)
- Most dangerous case: Strongly parasitic. The layered application uses some of the underlying operation’s resources to compete with the underlying operation, directly or indirectly.
I guess I wonder whether layered applications of the third variety have predictably long lifespans. If, for example, you’re providing a service that appears to sit at an online retailer and tells people how they can use what the retailer sells without paying for it…well, doesn’t the retailer have some motivation to find ways to prevent your use of their resources? And aren’t they justified in doing so?
This is just musing. Maybe the online sites that become underlying operations for layered applications are run by such powerful and/or benevolent corporations that they would never worry about parasitism.
Then again, maybe not.