From a thread in another forum: Could CNN.com be published from Moveable Type?  How about News.com?  Would the interface support what these sites are doing?  Could you scale it?

If you have an opinion, post it.

Deane

MT would not scale to the size of CNN or News.com. You could *probably* do most of what they are doing as far as end-user interface, but as far as sheer number of pieces of content, MT can't scale that far.

Whatever do you mean by that??

I think the question of whether it could scale depends on how the pages are set up.  If the MT output is very basic (just DIVs with classes), and all the content surrounding it was from include files processed by the Web server at request time, I don't see why it couldn't work.

How segmented can the rebuild process be?  I think you can rebuild just one file, can't you?  If the output was bare enough to be insulated from template changes then you may never have to rebuild for formatting reasons, just for content changes.

How about user interface and functionality?  Is there anything CNN is doing that couldn't be supported by the system?

Deane

Once an article is entered, it becomes a static page, so what's it matter?  :)  The ids go up to 999999.  I think I remember a thread about how to hack in another digit, too.

You can set the number of entries to rebuild with EntriesPerRebuild in mt.cfg, so if your limits are low on the server I'd guess you could lower how many entries to do.  Plus, you can cut down on rebuilding using includes.  A full rebuild isn't required very often, anyway.

But that's just my thoughts, it may be more technical than that.  :)

Well, you could publish just about anything with MT but not neccessarily well.  What it really comes down to is choosing the right tool for the job, and currently, MT is far from the right tool.

Yes, MT could handle the input and perhaps the output, but it would be a lot like trying to juggle chainsaws in a small box.  Currently, MT doesn't have any of the sorts of workflow processing  features that large (expensive) CMSs feature.  There basically can be no editorial oversight.  There is no automatic cross-referencing of articles, the keyword-based metadata is primitive, blah blah blah.  And don't get me started on trying to manage templates for a site that big.

Even small zines have to modify MT to suit its needs right now.  I'm sure that that won't be the case when this summer rolls around, but for now, CNN would be foolish to even consider it.

Before you go dreaming, let Ben and Mena focus on getting MT to the point where Wired could use it.

I mean quite simply MT couldn't scale that far. As it currently stands, there is no way to rebuild a group of entries without rebuilding all entries. fooljay's points are quite valid.

It's not like this is a big deal - MT isn't meant to be a commercial CMS on the scale of Vignette or something on that line. It isn't meant to run hundreds of thousands of pages across a server farm. The big content management solutions are in and of themselves almost operating systems (and in fact some are) that are designed to run mega sites across hundreds or thousands of servers. MT simply isn't. I love MT for what it is - a blogging tool, a CMS for a small to medium site - but I'm also realistic about what it's not. I run a site (tubescan.com) where we use MT as a CMS, and quite frankly, I've had to come up with pretty ingenious workarounds or simply do without for features that MT is lacking as a full-fledged CMS. And who's kidding who - tubescan.com is not a big site. We get decent traffic but we are not even in the same stratosphere as CNN or news.com.

I think Wired could use it now.  So could News.com.  Perhaps MTPro will bring us closer to CNN-like functionality.

Deane

hello! EntriesPerRebuild.

(and fooljay makes good points)

Psst. I don't think that's quite what he means.  That still rebuilds all entries.

oh yep, true!  thanks.    :)

I'm just thinking that the rebuilding part isn't that big of an issue.  

I'm with you on that. Trying to figure out what needing to only rebuild certain entries has to do with CNN.

With the CNN thing, I think right now the biggest drawback would be not having subcategories.

Perhaps we are using different connotations of 'could'.  You seem to be asking whether they technically could do it.  The answer plainly is yes.

So plainly in fact, that I moved on to the more interesting question: could they do it and stay in business.  My answer to this question is no.  Or at least I highly doubt it for the reasons stated above.

Nope, I could hack in subcategories in no time.  The biggest problem is that there are only two entry states right now: Draft and publish.  That would be a major, major problem for most publications, much less one the size of CNN.

For instance, in this system, ANYONE with a user account can post.  Can you imagine the CNN editors going along with the "honor system"?  No chance.

I've used some of the big content management systems (Documentum, in particular).  By and large, they suck.  They're massive, bloated, soul-sucking platforms, and only about 10% of their functionality is usually successfully implemented (or needed).

If MT published to a staging server, and there was a process on that server to distribute off to a server farm or distribution network (Akami, et. al.) then I think it would awfully scalable.

But, I see your point.  Workflow would be a big hole (though this is promised for MTPro).  However, understand that the big CMS products can be far, far more trouble that they're worth, coming from someone who was the project manager for just such an implementation not more than four months ago.  I can get more done with MT and 10 minutes than I got done with Documentum and 12 months (an exaggeration, of course...).

I'd be happy if we could find an MT installation for a site as professional, well-traveled, and consumer-accepted as Wired or News.com.

Deane

I've used Documentum and StoryServer (Vignette) before.  I completely agree.  But I didn't say that they were great.  I only said that they do the job editorially.  That is, there are roles and levels of permission.  Without those, I can't see any "respectable media outlet" going for it.



This is precisely how it would have to work if they were to use MT.  There would have to be separate MT installs, or at bare minimum separate blogs, one of which was internal and one that's external.  Editors would do the dirty work of copying approved and edited stories over and hitting publish.

Oh, also we're forgetting about the massive amount of visual and audio assets that CNN is managing.  That I wouldn't want to get NEAR MT....



But that's not the point.  One person doesn't need Documentum.  A thousand people very well might.  MT is much faster to set up and develop, but over the long run in a large publication, you'd put all of those hours back into the system doing production backflips...



Totally agree.  And you will, I bet.

Judging from what I see companies like Oracle are doing with their application servers, basically making more CMS type functionality - XML for instance - as part of their core product then the less likely it is that the Vignettes' of this world have a rosy future.

Anyone interested in seeing what a more full-featured content management system complete with roles and permissions, etc. might want to take a look at the open source project called 'Plone'.  

http://www.plone.org , which is itself a site that shows what a Plone installation looks like "out of the box".  You can log in and create your own page/site at www.plone.org, grant permissions, etc.

Plone is built using the open source web application framework, Zope.  It's Python based, but it's possible to use other scripting languages within it.  

Understanding how Plone and Zope work has been said to require a "major mind tweak".  It's all beautifully object oriented using Python's OO features.  Plone works okay for a lot of purposes "out of the box" so you don't really need to get into inner workings of Zope or do any programming, just get to understand how the framework fits together.

Zope is used for some pretty big sites, e.g., the New York CBS affiliate: http://www.cbsnewyork.com/

I considered using Zope or Plone for a new site I'm about to develop for my own small company.  But I'm pretty sure I'm going to do it with MT, since MT has most everything that I want and doesn't complicate things with a lot of stuff I don't want or need.

Regards,

Herb

I've toyed with Zope a little bit.  It's kind of a religion, in a way.  Deeply conceptual.  "Mind tweak" is a good way of putting it -- you kind of have to throw out everything you know about Web development before you get started.

Deane

What if you change a property that only affects a certain group of entries (i.e. a topic name)? Do you really think they're going to sit there and wait for x thousand or hundred thousand articles to rebuild just to fix/change something on a few?

Granted, most of the issues mentioned could be hacked in. But as an off-the-shelf product, it doesn't boast a lot of the features that the "big guys" (no matter how flawed) offer.

I just don't think that given the way MT is currently built it is realistic to expect it to be able to scale that far. Between the rebuilding issues, and workflow holes, and a few other things (subcategories, author permissions, templates) I just don't see it happening.

And finally, there is a stigma associated with low-priced products among big companies. I work for Jelsoft who make vBulletin and a lot of companies won't even look at us cause our highest price is $160 vs UBB which has a price scale which extends to thousands of dollars per year (for much less functionality). Companies assume that low-priced products must have something wrong with them or they just can't scale.

Oh, I agree and it's sad.  There is too much prejudice against Open Source code out there.  The party line states that they want a support number to call if there's a problem, but is it just coincidental that these numbers get used the most on crappy commercial products that cost six figures?

I would love to see MT get picked up by a large-scale commercial organization for their site.  I posted in another forum, asking what the biggest MT install out there was.  The only response I got was the MT Plugins directory.

Deane

Actually, that response was my personal blogs  kadyellebee and lovesongs and lovelinks and my daily to do blog and my photoblog, and probably a few more I'm forgetting, along with TheRedKitchen.  The mt-plugins.org site isn't nearly that big yet!!!

Kristine

May I ask why?  I mean I love MT and I enjoy stretching its functionality and all, but why would you want to use it to do cnn.com or something of that scope?  There are better suited tools that are designed for sites like that (and yes, some of them are OpenSource).  Trying to encompass that functionality in MT would make it unnecessarily complex for it's stated purpose:

The complexity of managing a site like cnn.com is orders of magnitude removed from what MT is intended for.  It would be like trying to cut down a forest with a hand saw.  You could do it, but using a chainsaw sure is much easier.  Around my house, on the other hand, a hand saw is just the right tool for me.

Good point.  Let me rephrase...

After working with enterprise content management systems for a number of years, I'd like to find an ECM that's as simple, user-friendly, stable, extensible, and cost-effective as MT.

And I'd like pigs to fly.

Deane

Here's a great article that relates to blogging in general, and this thread in particular:

Blogs as Disruptive Tech
How weblogs are flying under the radar of the Content Management Giants

Consider this:


I could see this happening.

Deane

This is a great discussion, and although I don't have much to weigh in with, I'm actually learning a lot about Movable Type and CMS in general.  This will definitely help me in preparing my new site for launch

If an MT frontend could combine with a phpBB ( http://www.phpbb.com ) backend with something like phpnuke to boot, it would be unstoppable.

ciaran