Summit must be the secular name our species came up with when we decided that certain business meetings had to have the same weight as religious conversions.

I’m not trying to be contentious by using this vaguely religious metaphor, so please consider me a radical centrist. But after a few months of vendor meetings for the analyst community in which each took us to the mountaintop to survey — via PowerPoint — a future vision for the valley below, I am almost all summitted out.

The religious reference struck me this week at the SAP Summit, the vendor’s very good analyst gathering in Boston, because, like some religious conversions, there seems to be a necessary pain component intended to make the conversion stick. (In most analyst meetings, the pain comes from sitting still for many hours of the aforementioned PowerPoint presentations.)

So, what did I learn? Well, lots — though I’m a CRM guy and much of what was proffered involved visions of a broader valley. The biggest impression I came away with was intramural since, having been to this year’s Oracle OpenWorld and Salesforce.com’s Dreamforce, I’m in a mood to compare, contrast, synthesize, and perhaps even prescribe.

Put everything I saw at these and other conclaves into a food processor, run it on high until something resembling peanut butter forms, and the result, to me, looks like this:

Software-as-a-service (SaaS) computing has won the battle, maybe even the war, but the victory is not enough to secure a homogeneous peace.

Translation: SaaS is important and remains the future of software, but there are multiple reasons why it will not reign supreme — not for a while at least.

[More, after the jump...]

A few weeks back, I noted that there are still some 6,600 mainframe computers — not only in existence but actually in use — and it will be some time before the population dwindles to the point that, like the B-24, there will only be a small handful of them capable of doing what they do. The same is likely for premises-based enterprise software.

It’s not that on-demand technologies can’t do everything that the premises-based products can. Rather it’s that the vendors of premises-based solutions have examined alternatives and decided that it’s enough for the moment to retrofit their wares with some of the best benefits of SaaS. It’s also because customers of certain application types are not ecstatic about sending those applications to the cloud.

Such has been the messaging that SAP, Oracle, Microsoft, and even Sage have been selling for some time now — and it has legs. No doubt this is a vendor-generated paradigm extension. Most important to this strategy has been the effort to retrofit products with some of the popular attributes of SaaS.

For a long time, things like better total cost of ownership, near-instant deployment, and great configurability characteristics were the exclusive province of on-demand solutions. But much of that has changed — now, enterprises that once looked longingly at SaaS solutions have some reason to put more gas in the old car and drive a while longer.

That was a strong message I got from SAP: No, it doesn’t — and won’t — offer every product in an on-demand format, but it can still deliver on the hot-button customer issues, to a degree.

I’m still puzzled by the customer who said he has 16 instances of an SAP system and that he schedules updates and maintenance for all of them. The SaaS-ist in me says, why?

To be sure, I don’t see any new applications coming to market that have not been conceived to be SaaS. That, alone, marks an interesting point in software evolution.

[Editors' Note: See Pombriant's take on evolution in his CRM magazine columns here (http://sn.im/0908rc - Evolution from Without," September 2008) and here (http://sn.im/0909rc - "Survival of the Relatively More Fit," September 2009).]

It’s quite possible that we are witnessing the bifurcation of the software business, and that part of the business will remain on-premises well into the future. That part will continue to diminish as a percentage of all software, especially as we see the continued expansion of platforms such as Salesforce.com’s Force.com — already more than 800 applications strong.

Perhaps this even means that, for a while, new applications will be smaller and focused on market segments rather than the all-encompassing front- and back-office systems that have occupied our attention for the last couple of decades. Twitter and Facebook, for example, have huge numbers of users but they are not nearly as complex as an enterprise resource planning system.

So my conclusion, after a lot of airports and hotel rooms — and as the circulation returns to my backside — is that we’ve achieved a kind of status quo between premises-based solutions and the cloud. The frontier will keep moving to the cloud(s), but there is life in the old paradigm.

I suspect this will be a good thing, as we turn some of our attention from software wars to the substantive question of how we do business after the bubble and its burst, as liquidity remains a serious challenge and, in the wake of Copenhagen, in a world more acutely conscious of sustainability.

Denis Pombriant, founder and managing principal of CRM market research firm and consultancy Beagle Research Group, has been writing about CRM since January 2000, and was the first analyst to specialize in on-demand computing. His 2004 white paper, “The New Garage,” laid out the blueprint for cloud computing. A CRM magazine columnist, he often guest-blogs with us at destinationCRMblog.com, but his own blog can be found here. (His Reality Check column on Marc Benioff appears in CRM’s November 2009 special issue on Salesforce.com.) He can be reached at denis@beagleresearch.com, or on Twitter (@denispombriant).