Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×
Announcements Content Engine Now Open Source 60

webword writes: "Did it slip under the radar that's content engine is now Open Source? The engine powers sites like,, and Not too shabby." More to the point, it powers Mr. Showbiz, one of my favorite movie sites. The license is based on the Apache License.
This discussion has been archived. No new comments can be posted. Content Engine Now Open Source

Comments Filter:
  • Respondants seem glad at this announcement because it means this software is now free (beer).

    This is inevitable, because the software has already been developed, and so there is no development benefit to be gained using the superior(?) open source methodology.

    It would be a bit more interesting for us open source agnostics if some company announced that it was planning to develop something new, as an open source project - not just dumping some huge source base for open source maintenance....

    'Cause I guess that I reckon Open Source is a lousy method to create good new software, but a decent way to maintain something, especially once you don't want to do so yourself.

  • to the converted.
    I doubt anybody here has to be told that open-source is growing.
    We all know that companies are jumping on the bandwagon.
    These pieces of info are really not that interesting anymore, maybe we should get this out to the windoze lusers somehow, it serves no purpose to prove to us that we are on the right side allready.
    Please don't bother me again until something Like Diablo II or Quake 3 or something has been gpl'd.
    If microcrud decides to open-source w2k - I think you could post it with some kina "major threat -Linux users unite and take up arms!" theme attached.
    Anything short of that being open sourced, I really am not that interested.

    PS. If 8000 people are, just say so, don't flame me because I feel different, I will democratically follow the crowd, no need to waste your napalm.

  • Keep in mind that Tea's function is not to determine an appropriately ranked list of sites using a user's search request. Instead, Tea takes the results from the search engine (or any other source of content) and uses templates to present them to the user.

    Just a friendly reminder to separate the apples from the oranges before passing judgment.

    - PatientZero

  • Think of what RedHat, Caldera, or whoever could do with a single package now. RedHat 7.0, webserver addition: includes your end to end large scale site solution, web services by Apache, Content Management by With a good ecommerce suite, that could be one hell of a package if it all came preconfigured.

    The stuff provides you with the tools to build a website - it'll never be an out of the box solution.

    CMP released their perl-based CMS a while back - Mason []. It looks very nice, it's used on some high-volume sites, and has a feature site suprisingly similar to Vignette's [] StoryServer (which is TCL-based, and grew out of C|Net []).

    None of which will do anything magical out of the box. recently had a decent article about CMS pros and cons, but I can't find it. Anyone got the URL?

  • How does this engine compare to other template environments, like webmacro, Novell's Internet Template Engine and others?
  • Wow, yet another totally incompatible servlet based templating system! What most people don't realize is that this technology is fundamentally 3 years old. Looking in the source code I find this: Author: Reece Wilton. If I then go to's who's who directory I find: Starwave Corporation Java developer (servlets, applets, beans, threads, etc). That's right, this is the same lame Starwave technology that they were pushing 3 years ago. Only now, instead of it being 100% proprietary, it's 90% proprietary, being somewhat Servlet conformant now. The analysis of the problem with servlets (Jason Hunter) is a very superficial reading into the problem. If it's simply a matter of needing a templating mechanism, then JSPs solve this in a standard way. But the bigger problem is that none of these technologies provide for a clear MVC architecture. For a solution that is years ahead of the GO crap in its thinking, look at the STRUTS project in the Jakarta initiative (Apache org). There are many other App servers that have solved both the templating problem and the MVC/flow control problems in a sensible way (Apple's WebObjects, ATG Dynamo). Did GO pay you guys to put up this link? Why did you bother?
  • GO is not open sourcing their search engine. In fact, they don't even really own the rights to their own search engine any more - Ultraseek was sold to Inktomi. (See

    What they are open sourcing is the content management software used for sites like Mr.Showbiz and and a few particular parts of the portal. Most of the portal is still built on an arcane, completely propreitary system, that could never feasibly be open-sourced!
  • JSP allows you to put code into the templates-- it does not mandate it. Developers who know what they're doing would never put programming elements in their JSP templates. It's quite easy to program JSPs such that the templates have only declarative tags in them. I think TEA offers no benefit at all.
  • This is all very interesting.

    It's not quite a full content management system, since it's missing workflow, but it seems to handle templating quite well. And since it's open source, there's a distinct chance that a decent workflow management system will be grafted on.

    The real advantage of a system like this is that...well, it's fairly powerful and it's free. If you want something with this kind of templating power, you're stuck using a very limited set of toolkits, including high-end stuff like Vignette Storyserver (which can run in the USD100,000's). If you're a smaller company or don't have a system that requires such blunderbuss tactics for content, this rocks the house.


  • THE Patrick Naughton??
  • Okay, so Kettle is win only.


    It's not required to write the templates:

    Editing of Tea templates can be done in any text editor, Kettle is the easiest editor to develop Tea templates in because it is integrated with the Tea language.

    It may be a bit friendlier, but, heck, every editor I've ever used for HTML has left me nervous enough to at least make me double-check and optimize my code by hand in vi or emacs (or notepad, or bbedit, or...). And if your artists complain, give 'em dreamweaver or something and build the template out of that.

    I'd really like to see a head-to-head comparison of TeaKettle/OpenCMS/Mason/WebMacro etc etc. Maybe I'll do it when I've finished Diablo II.


  • Check out OpenCMS under Available for some weeks already. Written in Java.


    OpenCms is written in 100% Java and based on J2EE standards. Java Servlets are used to deliver HTML content created from XML templates. The whole content of the system is stored in a JDBC - compliant database. Java Applets or ActiveX - Controls are used as client side editors. For the basic components, well known Open Source components have been selected, please see our Credits page.

    So this may be the solution for the people, who don't like another template engine and win-only Editor (OK OK, the HTML component of opencms is ActiveX for IE, but with reduced funcionality opencms works in a simple browser).

    btw., is there no HTML editing applet available like the programmers of opencms state? They would like to use oe, but didn't find anything despite the ActiveX thing.

    Bye egghat.
  • I am a professional web developer who HAS given Zope a roll and am very unhappy with the results. It is not ready for prime time! And for my application, it is like trying to kill a gnat with an Uzi.
  • recently had a decent article about CMS pros and cons, but I can't find it. Anyone got the URL?

    Don't know about the article, but CamWorld has a CMS comparison table. []

  • It's a thorny problem and the solutions available now are very expensive and I hear not that great. This coupled with Apache and the free XML tools that are out there now should combine to make a killer, free, web suite a reality.

    If you have Apache and XML (and any other programming language) making a content management suite is not a complicated problem. The only problem is how to integrate it with a client's exisitng publishing process - which may be different for every client. But the code running it is not complicated - so it is easy to cutomize for multiple clients.

    Taking a look at Tea, I see a few things happening:
    1. They have named it after a househould beverage and used childish analogies to describe it's functionality. Surefire winner.

    2. The documents supporting it emphasize it's complexity rather than it's simplicity. This tells me they are trying to market it to people who will not be maintaining their own sites. To quote from another post:

      The text below is from the TeaTemplateLanguage.pdf file (240k). This information does not seem to be on any of the site's pages (requires a bit more digging), so I think it makes sense to bring up to the surface a bit here...

      Why use Tea?

      In general, neither developers nor page designers author Tea templates. The goal is that they be written and maintained by technical producers who are liaisons between developers and designers.

      If I'm developing a content management system for my site, I don't want a matrix of developer::designer::technical-producers who control the site. I want the people who develop the content to control the site and will build the system accordingly.

    3. Coffee and Tea.
      Usually I think one thing when people start talking Java and Web development. Something about being born yesterday. The corporate types like Java because it has a name they can remember, so they say it back to the system analysts whenever a new design team is in the office. The web sites that would benefit from the use of Java development environment are few and far between. With Java you'll get a long development cycle, expensive developers, expensive hardware, expensive support contracts, and preformance increases that aren't even seen as necessary by the Media Metrix Top 50.
  • I work (contract) at, and must make a few corrections:

    First off, itself does not use this technology. As others have said, it was developed by Starwave, which is now part of the GO Network (i.e. the corporation), and their technology is (as the article said) used by,, etc. Just not by the portal part of the business (at least, not much so far); we've got our own unrelated portal technology. This is due to historical reasons, since originally this was several companies, now merged.

    Secondly, the GO (formerly infoseek) search engine is unrelated to this story; it is a separate technology and it is not open source. (See news stories about the spin-off of Ultraseek, btw).

    Thirdly, I haven't become expert in Tea personally, but it's been suggested for certain future projects at itself, since other parts of the company like it. If you know of something that does the same things but better, as some posters implied, by all means let us all know, so we'll know not to waste our time on Tea. ;-)

    Lastly, we support a super-light version of the portal for internet cell phones, check it out at (or just "", which is fast to type; it'll redirect your phone to Also a mildly-light site for slow speed wireless devices like Journadas; see And see for info about a version of our site targetted to wireless Palms, e.g. Palm VII. (None of which use Tea at this point in time.)

    (The latter paragraph is a touch off topic, but those are the projects I work on, so... :-)

  • They claim to have released a complete system as open source but the code for the compiler/ide (kettle), which run only under windows, is not on the site. From reading the docs it looks like a well designed template language but the compiler/ide is completely necessary to make everything go.

    I think I will stick to Webmacro [] for now.

  • Yes, it's unusual enough for a Big Company to open source something, but the really bizarre thing is that this is *Disney*! Giving something away! When was the last time you got *anything* free from Disney?

  • 1) Neither Apple WebObjects nor ASP solve the workflow-related problems that Jason describes
    2) Putting up as Open Source the engine that is responsible for serving up millions of pages per day is always news. Saying that Go.Com paid to have this put up is ridiculous.
    3) Asserting that this is useless is even more ridiculous. This has obviously filled a huge need for Go/partners, and can fill this need for others. For people for whom this does not, they can adapt the code to their needs. This is what Open Source is about.

    Anyone who works with a professional graphic design staff will quickly realize that JSP/ASP/WebObjects have the most convoluted workflows and will either write a tool of their own, or find a tool like this and modify it to meet their needs.

    If I were to ever show JSP code to my graphic design team, expecting them to use it, they'd quickly throw me out the window. [literally. :)]

    I do agree that there should be some sort of standard template/macro language. Unfortunately, due to the wide variability in workflows amongst web groups, the limited publication of existing ones, and the diversity of platforms that people use, creation of one is very difficult. That is why it hasn't happened yet.

    Someone ought to take the best features from the existing systems and make a set of uniform standards at each step in the process. That has not been done yet, as far as I know. Doing this would enable people to quickly adapt their systems to the standards, plug generic standards-based extensions into their systems, and unify the number of standards that exist.

    I have to say, however, that I am always amazed that it's _huge news_ when someone posts yet another lame windowmanager because it's written by one of the "good guys" yet when a large corporation posts code, even under an 'official Open Source license,' not only is it ignored, but people actually complain about it and start whipping out conspiracy theories.

    Grow up, kiddies. Open Source will never work with that attitude.
  • I've used Tea. I use Notepad to edit tea files. You can use any text editor to edit your .tea files.
  • Who is Richard M. Stallman? I know his name sounds familiar but I have no clue who he is.
  • I'm sure Rob will be taking a good look at the Go Networks code to see what of it (if any) would be useful for the site.
  • Gee, I guess my sense of humour was above you - so sorry.

  • by Anonymous Coward
    Why did Go opensource their product? Good on them, btw, but are they hoping that the PR will outstrip whatever licensing fees they used to garner? Was business not doing so hot? Was this an act of pure geekish altruism?

    junk, and stuff. or whatever...

  • by Anonymous Coward on Sunday July 23, 2000 @10:26PM (#911077)
    Is tea compatible with java? I guess it is if you have the right kettle. :)
  • The GO publishing system works very well. If you have ever needed a way to seperate code and content, this is it.

    The Tea templating system can be picked up in a few hours. From there, managing a web site becomes extremely fast, simple and reliable.

  • Tea likely results from the fact that Go/Infoseek had to think about this type of problem before many of the tolls out there were mature.

    NIH syndrome is quite prominent in the web-development field - there isn't one tool that has clearly established dominance, and most shops still roll their own. I suspect this will change - mod_perl and PHP do have a fair following. JSP will likely catch on if Sun can get the performance up.

  • by Robin Lionheart ( 14795 ) on Sunday July 23, 2000 @10:36PM (#911080) Homepage
    > Is there a site out there that has thorough and *unbiased* comparisons of the different search engines out there?

    Search Engine Watch [] is a monthly newsletter which does in-depth comparison reviews of search engines.

  • An open source CRM is already out there (or at least the tools to build one): []

  • Hehe. Reminds me of the old saying about 'scratching an itch'

  • by Foaf ( 1882 ) on Sunday July 23, 2000 @10:52PM (#911083) Homepage
    TEA is a templating system similar to WebMacro. As I understand it, TEA keeps the Java code completely off the HTML unlike JSP which is a mix of HTML, custom tags and Java code.

    Check out this article The Problem with JSP [] for more reasoning behind it.

    BTW, cat-ing your code to a file is for girls. Real programmers use copy con program.exe

  • Hmm... Yeah most of the content management stuff out there is pretty crappy and expensive. Not a good combination.

    Has anyone actually used Tea and/or Kettle? I wish I could say it was a great package, but I just haven't had time to look at it.

    But free (as in speech) is good. That means that if it is crappy, someone can make it uncrappy. :)

    This is a super important piece of software for a major website to have available for free. Sure you can serve up a zillion hits with Apache, but how do you manage the pipeline of getting structured content from writers and artists onto that site in a timely and efficient manner?

    Microsoft Office 2000? ;-)

  • It is nice to see that "Open Sourcing" does not automatically mean the GPL. For those of us who beleive in the GPL's forced-redistribution, this is a pleasent change.
  • Content management is one of the hardest things to design when dealing with non-technical content creators. It is even harder to isolate the content creators 100% from the web designers, or application developers. Many slashdotters who run smaller sites may not appreciate the value of a content management system such as this. I have already started to go through the documentation in order to see if this is something I may be able to use.

    Even if I do not implement this, the documentation alone can be a great learning guide in how to properly create a large scale web site. As I am reading through, I am finding I have already implemented many of the ideas in my php applications, but not to the extent they may be talking about. The lessons talked about in the documentation are things many of us may need to tackle down the road. Using many of the ideas GO relies upon, a site is much more scalable than without them. Once you get 10 people from different backgrounds working on a site, which started out as just a single "web developer" deal, the rules change. If you site already conforms to the newfound requirements, all the better.

    I saw one comment about how Kettle, the template creation IDE only supports windows, and is therefore useless. Maybe so, but since it is open source, I be an effort could be made to port to a Linux, or cross platform Java version. Most shops big enough to make "real" use of this would have Windows on most PC's anyway (Thats just a fact of life.) A "real" content management system, and I'm not sure Zope really cuts the mustard (but I don't know), is something Linux in general has been lacking.

    Think of what RedHat, Caldera, or whoever could do with a single package now. RedHat 7.0, webserver addition: includes your end to end large scale site solution, web services by Apache, Content Management by With a good ecommerce suite, that could be one hell of a package if it all came preconfigured.

    While this doesn't have the slashdot reader appeal of Mozilla for instance, this is a VERY BIG THING for professional web developers.

  • A "real" content management system, and I'm not sure Zope really cuts the mustard (but I don't know), is something Linux in general has been lacking.

    It might be worth your while to actually give Zope a roll, as it seems to have *more* features than the stuff. Not sure why you haven't already, as you seem very upbeat about content management systems.
  • On the very lowest end, we have ASP...

    Why do you immediately dismiss ASP as a low end solution?

    Many large sites use ASP: I don't think that BUY.COM is a 'low-end' website.

    You may not like ASP, but please don't make a blanket statement like that. ASP is very useful for many people.
  • by steveh ( 10334 ) on Monday July 24, 2000 @01:21AM (#911090)
    Personally I would hesitate to call this a content management system. This is little more than a Java alternative to PHP.

    Separating content and presentation through templating is just one aspect of content management. There is also defining workflow for how items of content should progress through the publishing process, applying version control over the content and managing a user community and their access to carry out actions on content.

    Few content management systems seem to be actual real applications that do all this. Most just give you the starting tools that you could use to build such a system if you wanted to.
  • Kinda makes Amazon's squabble over one-click seem petty, doesn't it?
    Oh, wait. It already is petty.
  • A ?real? content management system, and I?m not sure Zope really cuts the mustard (but I don?t know), is something Linux in general has been lacking.
    I successfully used zope to do content management with some html dummies who were allegedly webmasters for their area. Basically, you just give them permission only to alter certain objects which are inlcluded in the pages. You can write stuff on top of that to do other functions. The ickiest thing is getting a dump of the site for use behind regular apache serving... Only because the zope code is rather intimidating for non-developers it was hard to write such a thing, and you will see the use of tools like wget recommended instead.

    Boss of nothin. Big deal.
    Son, go get daddy's hard plastic eyes.
  • Scheme is already powerful, efficient, and (in the context of web templates) easy. I've used the Kawa Scheme compiler to create a template system that allows for easy integration with Java objects, and is clearly the easiest template engine to use for SQL integration. Thanks to the kawa scheme compiler, templates are compiled directly to JVM bytecodes for speed.

    See [] for details and examples.

  • Well, since this Tea thing looks somewhat similar (although far more complex) than a project I've been working on, I figured I'd see if anyone had any opinions on my idea. I have a rather crappy site (with no actual releases yet (unless of course there's lots of interest)) here: [].

    The basic idea is simple templates and PHP modules that come together to help make building database-driven websites a little easier. Let me know what you think. I've recently been slacking off on the idea since I'm not sure how original or good it is. When I saw this story here, I figured it was a great time to see if anyone was actually interested in something like this. It seems Tea uses Java (which I've always thought of as big and slow). My creation (TARP) uses raw PHP4.0 as the solution.

    Anyway, thanks for looking. :-)
  • i can't believe that people think this was a search engine. it's not.

    it's a content publishing system.

    it really only makes sense if you have a more complex site, with a few technical people who deal with getting data into databases and a bunch of non-technical people who want to create lots of html pages without having to worry about all the technical details. most importantly, and what differentiates it from jsp's and asp's, it makes it impossible for the non-techy content person to screw up the database, or do much other collateral damage.

    oh, and kettle IS written in java, but they made the HUGE mistake of basing it on microsoft's WFC (windows foundation classes for java) because sun's Swing was so damn slow when the started on kettle. since microsoft is undoubtably dumping WFC (more likely porting it to C#) and completely dumping support for java, this SERIOUSLY needs to get open sourced to someone can port it to Swing, which these days, (java2 1.3) is plenty fast enough if you write good code.

  • This gets me on a topic I've been struggling with for, well, years now. What's the best way to implement a CMS?

    I was first weaned on StoryServer, and later Prism (CNet's internal version), and left that with the conviction that introducing code into a page is little better than putting page elements in code, and usually is worse. For anything of appreciable size, the code will vastly outweigh the HTML. So your pages are mostly code, unless you build your own wall between functionality and presentation. And even if you do, you're still working in an environment designed for display. (In the case of StoryServer, this means working in Tcl, which is an unholy mess for anything but trivial functionality. PHP looks a lot better as a language, though.)

    Right now I'm playing with Python-based XML-DOM stuff, based on the model used by Enhydra. Basically, you have a designer make a page with sample content, then you tag the HTML tags, read the whole thing into XML, and your program modifies, clones, and rearranges the pieces of the page as content dictates. This seems very top-heavy to me right now, and it often doesn't work as cleanly as I might hope, but it does allow designers to work completely apart from programmers.

    So I guess my point is toward those who say "but Go's stuff is 3 years old" and asking what, then, is the more modern paradigm.
  • Here I go to start my own content search engine... wish me luck!

    Hmm... I wonder what sort of hard I would need to get a good one going... something like the google clustor?

    Anywho... this looks good to me!

  • I for one was very dissapointed when Infoseek switched over to GO, and the search results were topped with payed spots -It seems to me to defeat the purpose of searching the web if paying site (i.e. commercial sites) are getting top ranking. I must say though, I am glad to see them opensource the engine... Now hopefully someone can make less commercially focused search sites, using their same engine.

  • by russcoon ( 34224 ) on Sunday July 23, 2000 @10:03PM (#911099) Homepage
    Don't get me wrong, I'm a professional web developer who uses myriad different languages in a given day, and while I won't hesitate to aplaud GO's move, I am not sure where TEA will fit. On the very lowest end, we have ASP, stepping up into the real world we have PHP and JSP, and for sites that need integrated content management, we have Zope, so my question is this, where does TEA go? Reading some of the docs, I like it's compiled nature, but what makes that different from JSP? Anyone used TEA? Thoughts?
    +-------------------------------------- ---------+
    | Sissies...real men program with `cat > file` |
  • by inkey string ( 35594 ) on Sunday July 23, 2000 @09:57PM (#911100) Homepage
    now to construct the ultimate pornography crawler... (hey, modify to your needs right?)
  • Is there a site out there that has thorough and *unbiased* comparisons of the different search engines out there?
  • ...disney sucks...think they can win us over with the open source! HA! they have another thing comming!


  • ...will it be until all of what we see is completely customized to the person looking at it? It is only a matter of time... they are able to superimpose adds onto tenis games... so when will those ads be targeted just to me? I kinda liked the days of good old information on the net... plain text baby!
  • heh heh couldn't resist, its my fav language--and apparently pretty well liked by corporate america as well. Now if only the damn gui's worked as well as the servlets...(and no, jdk1.3's client hotspot did NOT fix the speed problem).
  • by legoboy ( 39651 ) on Sunday July 23, 2000 @10:58PM (#911105)
    "If you open source it, they will come"

    And with an article about fraudulent porn sites only a few spots farther down the main page, I have been inspired to most respectfully suggest that yes - the news of anything being open sourced will indeed lead to a not insignificant number of enthusiasts shuddering orgasmically.

    (Peeve - Come being spelled cum, much like love is spelled luv)


  • A quick looks seems to say it's their implementation of a scripting language, so that would put it on par with JSP, as it requires servlet support.

    I don't know if it's better than JSP - have to have a go and see I suppose. And coincedently I have a free weekend coming up too! :)
  • This could be a very good thing. Just think if apache or some other web server came with a high powered search included. Making local web site searches could become MUCH easier. If a nice front end configuration tool was written and also included it could help a lot of sites with making there searching better. There is nothing more annoying that searching a site for something and having the search come up empty and then finding it on the site 2 minutes later. I know it is a bit of over kill for a small site but for a medium to large site it could prove useful. Just my $.02.
  • by Matt2000 ( 29624 ) on Sunday July 23, 2000 @10:08PM (#911108) Homepage
    I think the comments posted so far are missing the point, this is not simply a search engine system being open sourced but a content management system.

    This is a super important piece of software for a major website to have available for free. Sure you can serve up a zillion hits with Apache, but how do you manage the pipeline of getting structured content from writers and artists onto that site in a timely and efficient manner?

    It's a thorny problem and the solutions available now are very expensive and I hear not that great. This coupled with Apache and the free XML tools that are out there now should combine to make a killer, free, web suite a reality. [] - Funny
  • First thing is... WOW... cool... *wipes drool from chin* Upon going over the documentation, I feel inundated by having to digest and regurgitate yet another scripting language ....then I question, WHY open source it? Maybe I'm talking out my respective ass here, but this (and most corp open sourcing) has got to be a PR move. But, I'm not going to scorn Go for open sourcing it's search engine. They made a good move, and this definantly helps all web developers and up and coming sites (whoo hoo... lets hold a contest to see who can do their search engine first) Seriously though, the only real draw for the corps to get into open sourcing is the PR that comes from hopping on the bandwagon. I mean, it's free advertising due to the 'credits' for the original script the mod'd engine is made from. So ya, that's better than a banner ad. And as far as eating into their profits, I think not, as TEA was pulled from another product Barista and technically, this should help them with their full product sales to enterprises wishing a more thorough engine for their search portal. As far as the IDE goes (Kettle) That's NOT open source. Which really bites since it's Windows only. Hopefully, Go's working on a port to the penguin.
  • Wow.. that is news. Licenses for the altavista engine start at about $20,000.. The engine could save you some bucks..
  • >>Check out this article...

    It was interesting to read this article. When my company started doing web development, the first thing that struck me was all those print statements to output HTML. (This was MS ASP, but bear with me). Pretty obviously this was not the way to go. It surprises me that Sun didn't look a little further ahead when making their imitation of MS' ASP + COM components. We could have got a new integrated architecture designed for web page generation.

    Just as the article dicusses re. JSP and JavaBeans, MS developers had long since started using COM components (written mostly in C++ or VB) with the ASP pages becoming just a simple interface to the browser. However, this still didn't solve the basic problem of integrating HTML with code.

    Our solution was to write a simple template function, using VB COM components. This enabled embedded variable tags in the HTML, which was stored as normal txt files in the server directory. The variables could be set by the various presentation/business logic components, and expanded by a standard routine, or replaced dynamically by raising "missing variable" events. The whole bit of code took about two days to develop, and saved vast quantities of developer time.

    We still use it, although it is now a complete architecture for invoking response objects through standard interfaces, which took a little more work, though not much, and see no particular need for professional packages which are finally appearing to solve the obvious problems with the whole JSP approach

  • by TheInternet ( 35082 ) on Sunday July 23, 2000 @11:38PM (#911112) Homepage Journal
    The text below is from the TeaTemplateLanguage.pdf [] file (240k). This information does not seem to be on any of the site's pages (requires a bit more digging), so I think it makes sense to bring up to the surface a bit here...

    Why use Tea?

    In general, neither developers nor page designers author Tea templates. The goal is that they be written and maintained by technical producers who are liaisons between developers and designers.

    Tea resulted from several years of experience with other web page building mechanisms. Most web-based applications start out with HTML tags embedded in code, whether it be C, Perl, or Java. This approach is adequate for small or first-time projects because it doesn't take very long to develop.


    Rather than embedding an existing language into something like an ASP or JSP, Tea is a language specially designed to meet the requirements of a templating system. It is safe, simple, efficient, and powerful.

    In one instance, Tea is integrated with a special servlet. This servlet gives Tea templates control over page building, while retaining strong ties to a back-end application written by a Java developer. While this servlet provides functionality similar to that of JSPs, Tea enforces correct model-view separation because of the intentional language limitations. Although this is also the suggested separation model in JSPs, it cannot be enforced.

    They also seem to put an emphasis on simplicity. The loops and output statements are really clean. The aforementioned PDF sheds a lot of light on this. Another interesting note:

    Tea templates compile into Java class files and execute just the same within a Java Virtual Machine. The Tea compiler does not generate intermediate Java files and hand them off to a Java compiler, but it instead generates the class files directly. Still, the resulting class files are as efficient as any handwritten Java code.

    Because no Java compiler is used, Tea can be distributed much more easily. All that's required is a Java2 runtime environment. Sun's JRE can be used, but the SDK isn't necessary.

    The bad thing is that the IDE only runs on Windows.

    - Scott

    Scott Stevenson

The trouble with computers is that they do what you tell them, not what you want. -- D. Cohen