Forum OpenACS Q&A: openACS vs Weblogic/Dynamo/Broadvision...

I'm looking for developers with personal experience biulding on one of the propietary platforms to provide some insight into cost of developing in ACS vs any of these well-known propietary toolkits.

For the sake of argument, I'd be curious what it would take just to build the existing community functionality of ACS with any of these products in terms of:

License:
Dev Cost:
Dev Time:
Hosting:
Maintenance:
Risk:

Any real-world examples or personal experiences are greatly appreciated!

In my personal experience, I watched Levis take a great website that was launched in under 3 months for less than 100K and replace it with a Blue Martini site that cost over 1M and after 6 months they abandoned the project.

Collapse
Posted by Rodger Donaldson on

The biggest risk with a lot of these companies is that they're going under with the end of the Stupid Web Boom. BroadVision are appearing constantly on FuckedCompany, slashing staff left, right and center. Prophecies of doom WRT the danger of being on a proprietary platform from a vendor who went bust are beginning to become horrifyingly real.

As far as my experience goes with Vignette:

License:
Depends on your setup, number of visitors, etc, but can easily run to millions of dollars. Hundreds of thousands at least.
Dev Cost:
Factor on a couple of weeks for VPS to come in and check out your architecture, plans, etc. VPS are not cheap - you won't get much change out of $2000/day/person, so tens of thousands for the early dev phase. Vignette developers are expensive - rates in London appear to be on the order of 100 pounds an hour, for example. Since Vignette partners have to pony up substantial sums for the privilege of their status, they will charge big time, too. And Vignette training for in-house staff is expensive.
Dev Time:
Varies widely. If you want to use their suite of pre-built stuff, which has until recently been very slender indeed, you'll want to budget a good six months, minimum. That's assuming the VPS guys you get are good and you have a good source for competant developers with a chunk of real world experience. Be warned, there's a lot of crap on the market in the Vignette development world. Projects can easily blow out to years. Be warned that many of their add-ons are horribly untested, and I've worked at a couple of sites that have had very, very bad experiences with being early adopters.
Hosting:
You'll probably want at least a couple of systems of your live setup, and seperate dev and content management systems. You'll need a reasonable amount of grunt to run them; something like a pair of E250s is a practical minimum for low volume sites, rising to a cluster of E450 class web boxes and a higher end DB server. V/Series gobbles memory. If you use VAF (App Framework), double everything. Also, beware: the support teams can become unhelpful if you sepearate systems with firewalls.
Maintenaince:
Costs big time, but you get a lot of major upgrades free for it. Their support organisation, like any, varies in expertise from people who are very good to ones who are very bad. You can't get on site support (AFAIK) without paying VPS.
Risk:
Not much. Vignette have a lot of big customers, so they aren't likely to go broke. If management to decide to go with Vignette, the best thing you can do is put a lot of time into selecting people to work with you. Don't accept Vignette's recommendations, since in some contries their preferred partners are useless; I've made a lot of money cleaning up broken sites by preferred partners.

Lest I sound excessively negative: Vignette's no worse than any other vendor of a toolkit masquerading as a product. It's expensive, it sucks, but they all do. Vignette will be around in some way next year. Just avoid the sharks that follow them around.

I asked a question about Dynamo on the web/db board a while
back. You can read the answers here:

http://www.arsdigita.com/bboard/q-and-a-fetch-msg?msg_id=00
0YzM

Collapse
Posted by Janine Ohmer on
Here's some real-world feedback on WebLogic...

I haven't been to the Consumer Reports website in a while.
While I was away, they redid the site in WebLogic.

Firstly, it's slow as molasses.  Secondly, the search now brings
up way too many matches (it used to give only one link per
matching article, not links for each matching piece).  And thirdly,
the login page no longer loads, meaning the entire site is
virtually useless (since it's mostly a subscriber-only site).

Now granted, all of these may be due to incompetence in the
implementors, not WebLogic itself.  But I think it's safe to say that
WebLogic is making things harder, not easier, considering
these guys did have a functioning site before so they do know
what they're doing to some degree.

Right now I'm just really annoyed... our microwave has
committed hari kari, I've lost the issue with the most recent
microwave reviews, and I can't get in to the website.  Looks like
I'm stuck paying a slightly outrageous $7.75 to use their faxback
service!!!!

Collapse
Posted by Jamie Ross on
I spend a couple years as a senior consultant for Vignette VPS so I can add a few comments to Rogers.  First I think his remarks are probably a fair assessment.  Vignette Storyserver worked pretty well but it helped to have a good Oracle DBA and Unix admin as it runs about 4-5 processes integrating all the template management, updating etc.  I liked the TCL API and it was well documented with examples so was not difficult to learn.  The lack of a ACS stype toolkit was frustrating but they built kits , just you had to pay for them which I thought wasnt necessary given the cost of the main package.  I built a couple sites with workflow, content mamagement , personalization etc in about 3 months each with a couple local developers.  The main issue is getting the requirements right, making decisions etc rather than the technology.  So Vignette works fine if you have at least 1-2 Million to spend and want the security of a company with lots of Fortune 500 accounts.  I suspect they will be acquired by IBM or Oracle in the future..  so they will be around.

I don't have experience with the new java based stuff other than my main assessment of all the web app server stuff is that it is more complicated to maintain and java is not a great language for web development.

cheers
Jamie