We all know what an AppStore is, right ? Like the Apple iPhone AppStore. You can:
- Browse, search, view descriptions of applications
- Download and automatically install applications
- Be notified when new versions of your applications are created.
- Purchase applications and be billed for them
- Lets start off and clear the air. We'd have to TRUST whomever ran the Appstore. That person would have to proactively encourage folks to put content up, and whilst many BP's would stop short of actual selling via the Appstore, might be interested in the PR and ability to manage eval and test copies. I certainly would. And lets face it -that person might actually charge a 15% (The rate Apple charges, I think) or so premium from the sale in order to fund the appStore. Some BP's dont trust other BP'sl.
Why not choose Lotus or IBM ? Well, its been 9 months since last Lotusphere, and we still dont have 8.5.1, nor air-cover (promised for years). We've had a wonderful 'keep the yellow bubble pleased' 'Lotus Knows' campaign, but that hasnt translated into actual Brand Recognition. If IBM were to do something, some cynicals BP might say, it'll be late, slow, buggy, badly implemented and give a horrble user experience. The actual IBM web site, for instance, is hardly an exercise in clarity.
No, IBM probably cant (thanks to their lawyers or their horrible internal processes, or their cross-tier product set or their management) do it. Certainly, IMHO, they couldnt do it well. So we need to find a trusted, knowledgeable BP other BP's and end users are aware of and can trust. This might be difficult.
- The Users would have to be aware of this. This might mean Lotus actually giving people this URL. It'd certainly take the heat off of Alan Lepofsky's list of third party applications. (IBM does actually have a Business Partner catalog but - guess what. No-one knows about it, it sucks, and we've had a sum total of five click-throughs in five years. ). This would have to actually be the go-to place for all domino admins, whether SMB or otherwise.
- It has to be SIMPLE. I've installed packages in the past that came with four pages of text. Hell, WE used to distribute four pages of text. It has to be a simple click-to-action installer that will download the application, and allow you to install it from template(s), create replicas, set ACL's, create relevant groups, create mail-in databases, and so forth. And be able to UNINSTALL all that crap too!
- The method of packaging the application therefore is paramount, and for this to have any chance to succeed, has to be simple, easy, and open-source, so that EVERYONE understands whats going on. Something easy enough for a corporate to use it internally in order to speed up their own application distributions, but yet complex enough to allow complex applications to be automatically updated.
- License Management. If this were the microsoft channel, we'd have so many customers buying our products that most of the time we wouldnt necessarily care about license management. If we got wind of someone badly breaking the rules, we could set our lawyers on them (assuming they have a decent IP structure in their country). However, we're all chasing (thanks to less customers) a diminishing notes market, and we all need to protect our IP. So a mechanism that, whilst doesnt create a huge burden, can prevent most folks from ripping off code all the time. Trust as a model works most of the time - I'd certainly recommend a light touch.
- Maintenance, upgrades. This mechanism would have to notify the end-user that their licenses are expiring on the eval stuff, notify them of new releases (and perhaps that whole 'new version' release cycle) and allow rollback.
- There are a lot of very trustworthy Lotus BP's out there who might be right to at least headline this.
- A lot of the tools for all these operations have already been developed. We, for instance, have a complete build/install/deploy tool and installation wizard. I'm not promising we can donate it, but this stuff does exist.
- Surely even IBM - slow, ponderous IBM, realises that the only way its going to keep its diminishing pool of users is to supply better services. Better services - because of their own limiited Domino development ability in ISSL - means business partners. Business Partners are crying out for a better way to connect to customers (something the BP program sucks at) and have the product to turn a customer experience around.
- Imagine that, for free, every single item on OpenNTF was automatically included in this AppStore without any effort.
So far, so very very easy, right ? A user-provisioning engine with some sort of authentication, profile managment and so forth. And of course an interface for folks to upload applications to.
In Apple iPhone land, the apps are packaged, uploaded to the Appstore and then downloaded and automatically install. All without compromising user security, and so forth. This is a lot harder in Domino land:
Phew. A pile of stuff in there and I'm only just starting. However, look on the bright side:
So. There's a few big ideas floating around here - nothing I see is insurmountable. And there's everything to play for. Discuss.