Seriously, those aren’t my words, but Stafford Masie’s (of Novell at the time) at the CITI conference just after the deal was announced:

…Y’know, we’re a Linux company, we do identity management, but we’re a Linux company. Identity management, there’s so much happening there to open source alot of the APIs, which we’ve already done, the only thing we haven’t open-sourced in the identity world is kinda our directory, and I can tell you what, we probably won’t, because again – the same reason alot of proprietary vendors wont take their big software and unwrap it, like I’ve always said- if you unwrap this baby its ugly, people will run away, ok, there’s certain proprietary software that you never want anyone to look at…

Confidence inspiring, no?

Relax. As Roy said, I’m not defending Microsoft. In fact, I never even mentioned Microsoft specifically. The only products I mentioned were from Novell. Beyond that, I wasn’t suggesting anyone go in any particular direction. You have to do what works best for you.

What I am suggesting, however, is that different places find value in different implementations regardless of distributor. I figured I was pretty clear in that when I wrote “I’m not advocating one over the other”, don’t you think?

To address your specific points, from what I gather, OpenLDAP and Kerberos are only parts of the entire identity management puzzle. How much experience do you have with those technologies? I, possibly incorrectly, was under the impression that a “home grown” implementation of a directory service stack and ancillary processes is a difficult thing to scale up compared to some existing commercial directory and identity services. I know Redhat(or Fedora) Directory Services, formerly Netscape’s Directory Service is around but there is no data on how it performs or scales. Can you span an OpenLDAP implementation over multiple sites with variable link speeds and have a multi master design without doing some serious hand holding?

If an organization has the talent to address those concerns and the design works for them, then by all means I’m not going to say they’re nuts for not going commercial.

Really, you just have to relax. Just because I disagree with something Roy or anyone else here has to say, doesn’t automatically mean I’m pro Microsoft.

These are the very textbook definition mature technologies, both the software and protocol. There are even *gasp* graphical tools to manage the combination.

Groupware? There’s tons. Do a search yourself or pay me, but stop with the disinformation. Genuflecting 5 times a day to Redmond is fine, just don’t insist that everyone else joins you.

As for Novell’s lock-in practices, I guess it depends on the product. Zen 10 has migrated to a directory agnostic framework where you don’t need eDirectory. Groupwise is still tied to eDirectory. If I had to guess why, I wouldn’t say lock in, I’d say that there’s a low reward to money/effort spent with separating it at this point. I have no experience with Identity Manager, so I don’t know.

I think really you can’t make a blanket statement for how any organization should operate with regards to IT. Some prefer to consolidate around a certain product line or company, some might be prepared to support “disparate” systems.

Well, it’s also risky to take a monolithic approach where everything a business uses is from IBM or Microsoft, which increases lock-in and dependency (hostage situation). Components don’t interact with anything but other components from the same vendor. I’m not too familiar of Novell’s lock-in practices.

You don’t need to dine at one place or rely on one support contract/supplier.

I think really you can’t make a blanket statement for how any organization should operate with regards to IT. Some prefer to consolidate around a certain product line or company, some might be prepared to support “disparate” systems. Regardless of what direction an organization decides to go, you can’t make the assumption(which it seems you’re doing) that all products in various market segments are equal.

To use an analogy, you don’t avoid dining at an Indian restaurant just because they have no egg roll.

@Needs Sunlight:

The problem is that neither Canonical or Red Hat have mature product lines in certain market segments like groupware or identity management.

http://www.betanews.com/article/New_Netflix_player_uses_Silverlight_to_reach_Mac_Linux/1225120899

If Novell forks a project, apparently for the purpose of infecting it with M$-licensed, proprietary technologies, the original project has no obligation to accept material from the contaminated fork.

Nor does the community around that original project have an obligation to audit material from a contaminated source. That leads easily to a denial of service where all the experts’ time is wasted with revisions, half-baked products, and even red herrings. Case in point:

http://arstechnica.com/news.ars/post/20080925-judge-microsoft-documentation-unfit-for-us-consumption.html

There’s never been an excuse for any company to be pushing WMV/WMA:

http://www.computerweekly.com/Articles/2008/02/27/229597/european-commission-fines-microsoft-680m.htm

Rather, any company pushing WMV/WMA needs to be dealt with for contributing to further damage to the European market. The Noorda’s are gone. Allison and his colleagues are gone. Only the name Novell remains and that appears to be a shell for marketing MS traps. Canonical and Red Hat are two better places to be steering business. Or check your local business district. Chances are good that there is already at least one well-established shop, with a good reputation and satisfied clients, that sell linux support. That is, if your in-house shop is unable or unwilling.

http://www.archive.org/details/e-dv034_01_vikram_dendi_001.ogg