Re: [Rife-users] A couple of questions before starting with RIFE
Hi Oliver, welcome on the list! (we already chatted a bit on IRC) I`currently working on my Bachelor Thesis in Computer Science and use RIFE as the Target Framework for an MDSD (Model Driven Software Development) Tool as an Eclipse Plugin. That's the interesting part ;-) The Thesis will contain a Eclipse Plug-In to model a web application and generate the application from the model. If i can get it the generation process will take care to do not overwrite the parts written by the developer. One Option to do this is by using Protected Areas, but that will be done in 1-2 weeks. Did you ever consider to not generate everything, but to evaluate your model at run-time, using a set of well defined APIs and hooks and a component model for customization and integration? That's the approach we took with RIFE/Crud. The advantage is that the developer never is able to accidently change something he shouldn't, simply because there's no generated code to change. His source repository is also not cluttered with a wealth of files he didn't write himself. I think that when you generate something from a visual schema, you should allow full roundtrip editing so that changes that are made to the sources are reflected in the UI. That's just my take on the topic ;-) I try to make it abstract to be used with other web framework (or even RoR). But at the moment i only support RIFE. The Version in my thesis will only contain the basic functionality, because lack of time. But after the thesis i will contribute the sources to RIFE to be extenden and hopefully be useful. The Thesis will end on 24.2.2006 and i will hopefully finish before the 15.2. After this i have to held my colloquium and then I`m finished. (If nothing goes wrong.) That's very short notice. The best of luck with your project and if you need suggestions, help or are stuck, don't hesitate to speak up. I believe so too; and people like Bruce Tate are singing it's praises... and if you already know java, you don't have a lot of reasons for not using Rife as-is now; and it will only get better :) correct :) Don't forget the conference talks. RIFE is really going places nowadays. Two talks at JavaPolis next week. Two talks at TheServerSide Symposium in March, and there's a good chance that I'll speak at JavaOne in May too (someone on the expert team contacted me to make sure a proposal was submitted for RIFE, since the were interested in the topic). Best regards, Geert -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Hi, ok to make it short. I introduce myself a bit. As mentioned in my previous mail. If have some years on java development, as a fulltime developer and as a student. My focus was web development and generators in a B2B Company. Worked with Struts for some time and wrapped it in a own framework, but that doesn`t make the things better. I`currently working on my Bachelor Thesis in Computer Science and use RIFE as the Target Framework for an MDSD (Model Driven Software Development) Tool as an Eclipse Plugin. The Thesis will contain a Eclipse Plug-In to model a web application and generate the application from the model. If i can get it the generation process will take care to do not overwrite the parts written by the developer. One Option to do this is by using Protected Areas, but that will be done in 1-2 weeks. I try to make it abstract to be used with other web framework (or even RoR). But at the moment i only support RIFE. The Version in my thesis will only contain the basic functionality, because lack of time. But after the thesis i will contribute the sources to RIFE to be extenden and hopefully be useful. The Thesis will end on 24.2.2006 and i will hopefully finish before the 15.2. After this i have to held my colloquium and then I`m finished. (If nothing goes wrong.) Geert did mention a book down the line; i for one won't mind contributing a few chapters to that endeavor. Maybe there`s place for an Model Editor Chapter ;) I think it`s better to grow slow and got instead of the quick hype that ends after two years. I believe so too; and people like Bruce Tate are singing it's praises... and if you already know java, you don't have a lot of reasons for not using Rife as-is now; and it will only get better :) correct :) Ok enough poor man english :) cu on irc, odo (_odo_ without the underscores is reserverd by someone else) PGP.sig Description: Signierter Teil der Nachricht ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Geert did mention a book down the line; i for one won't mind contributing a few chapters to that endeavor. Cool Emmanuel, I'll keep that in mind. I postponed the book proposal for a couple of months since I'm swamped with work at the moment and the new-year festivities are coming soon. In January or February I'll resume working on the TOC. I'll ping you then. -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
RIFE needs to get a bigger contributors/developer base. Maybe a book, some articles in the press. I think it`s better to grow slow and got instead of the quick hype that ends after two years. I have a guaranteed ticket for being published on onjava.com and theserverside.com with RIFE articles. The only thing is that I always get stuck when writing them. I started one a few months ago, but I fear it's too dry and people will give up on reading after the initial paragraphs. Also, the example is quite boring: http://uwyn.com/rife_tss/rife_tss.html The main problem is that I think it's important that people understand the site-structure and declaration from the start, but it's also the most boring and non-hype topic. So I think I will start over and begin with RIFE/Crud, introduce the CMF, customize things with properties, add new elements, gradually introduce site-structure concepts and talk about testing. Just to put the 'wow' factor as much in the beginning as possible. Any ideas for a cool, short example application that would fit the bill? Preferably one with logic and data flows. -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
> RIFE needs to get a bigger contributors/developer base. Maybe a book, > some articles in the press. I don't know how many people are quiet on the list (there are a few lists I belong to that I've never posted a single msg to) but I think, at least for the last month or so, people are slowly adopting it... I believe it will get its due of developer mindshare in time :) I like the fact that the documentation is open and anyone can add to it; If you are using it now, register yourself on the wiki. If you find something (you couldn't initally find on the wiki) add it in--it helps others down the line, and reinforces what you've learnt Geert did mention a book down the line; i for one won't mind contributing a few chapters to that endeavor. > I think it`s better to grow slow and got > instead of the quick hype that ends after two years. I believe so too; and people like Bruce Tate are singing it's praises... and if you already know java, you don't have a lot of reasons for not using Rife as-is now; and it will only get better :) > Writing my Thesis about a Eclipse Plugin. It`s a Model Editor that > generates a RIFE app, but i will write a Mail later ;) very nice... pls, do share :) -- eokyere ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Hi, a short answer from a RIFE Newbie :) 1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? Don`t know why it is not wider adopted, i think it is not because it`s from belgium. Many developers looking for alternative web frameworks these days. RoR got this hype because it has the quickest startup you can get. (RIFE is as quick as RoR with RIFE/ Jumpstart as I later learned). A lot of developers left the java trail and used the quick and dirty ones. A lot of them have bad experiences with Struts and other big frameworks making it hard to earn your money. But that`s the same we have seen for some years with PHP, Holy Grail of Web Development. RIFE will come slow but it will. It`s very interesting and don`t stand in your way. It`s not easy to learn, the docs are old at some places but the cookbook is fine. Support on IRC and List are excellent. RoR as an examples has the advantage that they have a written a big book with everything explained, without this a lot people won`t adopt so quick to RoR. RIFE needs to get a bigger contributors/developer base. Maybe a book, some articles in the press. I think it`s better to grow slow and got instead of the quick hype that ends after two years. 2. How long do you think it would take a reasonably experienced Java developer, who has worked mainly with the Struts/Spring/ Hibernate approach to web apps, to get comfortable and productive with RIFE's very different way of thinking/working? You can learn it really quick. Just stop trying to put the logic in the JSP. If started with it and loved it because i know if i start the next project with a team i don`t have to worry about a lot of things. Clean seperation of thins, features i need and growing. 3. Are there any tips anyone could give me about how to start 'thinking in RIFE'? I.e., how to approach the design of a web application with RIFE in mind. Perhaps something along the lines of planning it as a flow diagram, thinking of possible exits from each page view? Writing my Thesis about a Eclipse Plugin. It`s a Model Editor that generates a RIFE app, but i will write a Mail later ;) cu odo PGP.sig Description: Signierter Teil der Nachricht ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] Implementing HTTP Basic (was: Tomcat) Authentication
Fred, you seem to have trouble to divide the architecture in its different modules. The authentication elements just provide one a specific interface to do authentication. This has got nothing to do with the Credentials, CredentialsManager, SessionsManager and SessionValidator in interfaces. You can use there manager classes to build any kind of authentication interface, including HTTP authentication. Forget about the concrete elements, and instead of using forms, request parameters, cookies, ... just implement an authentication element that does what HTTP authentication needs and use the interfaces above for the backend. On 5-dec-05, at 19:40, F Baube wrote: Surely sendmail reeled when thusly spake Geert Bevin: Fred, what is important for you, to have HTTP Authentication or to access tomcat's database? That is indeed precisely the right question to ask :) The short answer is: HTTP Authentication. The longish answer is: It appears that Rife is architected to be incompatible with HTTP Basic authentication ... BUT this is easily fixed. I would implement HTTP Authentication using RIFE's RoleUser credentials and authentication session managers. This gives you container independence and easy migration in case you need it. I guess then you are referring to the instructions given at http://rifers.org/wiki/display/RIFE/Flexible+authentication The relevant classes appear to be com.uwyn.rife.authentication.credentials.RoleUser com.uwyn.rife.authentication.elements.Authenticated com.uwyn.rife.authentication.elements.RoleUserAuthenticated and possibly also DatabaseAuthenticated, MemoryAuthenticated, MixedAuthenticated It seems tho that actually the existing authentication classes could be lightly modified to conform to HTTP Basic authentication. Currently, when authentication fails, Rife replies with an HTTP response "200 OK" (!) This will _not_ trigger the appropriate behavior in client apps (like browsers !) which look for a "401 Unauthorized" and, if it is found, generate a login dialog. For example, a standalone WebDAV-capable editor application knows to display a login dialog when it gets a "401", and it feels free to ignore the HTML response body. This is also how Tomcat's "manager" servlet operates. I would suggest that class Authenticated could be given a new boolean configuration parameter (say, "httpbasic") which would direct the Element to (a) respond to login failure with "401 Unauthorized" (plus an [optional?] login template), and (b) set the childTrigger if (and only if) the "Authorization:" response header is valid. Does this sound correct ? fred -- F.Baube* Georgetown/MSFS/1988 * Act locally. email fbaube#welho.com * Think pangalactically. gsm +358 41 536 8192 * wmd 60°11'10.8"N 24°57'36.9"E ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
[Rife-users] Implementing HTTP Basic (was: Tomcat) Authentication
Surely sendmail reeled when thusly spake Geert Bevin: > > Fred, what is important for you, to have HTTP Authentication or to > access tomcat's database? That is indeed precisely the right question to ask :) The short answer is: HTTP Authentication. The longish answer is: It appears that Rife is architected to be incompatible with HTTP Basic authentication ... BUT this is easily fixed. > I would implement HTTP Authentication using RIFE's RoleUser > credentials and authentication session managers. This gives you > container independence and easy migration in case you need it. I guess then you are referring to the instructions given at http://rifers.org/wiki/display/RIFE/Flexible+authentication The relevant classes appear to be com.uwyn.rife.authentication.credentials.RoleUser com.uwyn.rife.authentication.elements.Authenticated com.uwyn.rife.authentication.elements.RoleUserAuthenticated and possibly also DatabaseAuthenticated, MemoryAuthenticated, MixedAuthenticated It seems tho that actually the existing authentication classes could be lightly modified to conform to HTTP Basic authentication. Currently, when authentication fails, Rife replies with an HTTP response "200 OK" (!) This will _not_ trigger the appropriate behavior in client apps (like browsers !) which look for a "401 Unauthorized" and, if it is found, generate a login dialog. For example, a standalone WebDAV-capable editor application knows to display a login dialog when it gets a "401", and it feels free to ignore the HTML response body. This is also how Tomcat's "manager" servlet operates. I would suggest that class Authenticated could be given a new boolean configuration parameter (say, "httpbasic") which would direct the Element to (a) respond to login failure with "401 Unauthorized" (plus an [optional?] login template), and (b) set the childTrigger if (and only if) the "Authorization:" response header is valid. Does this sound correct ? fred -- F.Baube* Georgetown/MSFS/1988 * Act locally. email fbaube#welho.com * Think pangalactically. gsm +358 41 536 8192 * wmd 60°11'10.8"N 24°57'36.9"E ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] Users guide as PDF?
Yup, just as Geert mentioned; i was going to point to http://confluence.atlassian.com/display/CONF14/Confluence+to+PDF cheers, -- eokyere On 12/5/05, Geert Bevin <[EMAIL PROTECTED]> wrote: > You can export anything in the wiki (included the live guide) from here: > http://rifers.org/wiki/spaces/exportspace.action?key=RIFE > > On 5-dec-05, at 17:29, John Moore wrote: > > > It would be handy, IMHO, if the users' guide were available as a > > (preferably printable) PDF. It may well be that Rife provides the > > facilities to do this already and it's simply a case of Geert > > pressing the appropriate button? > > > > If anyone has generated a PDF from the HTML and could e-mail me it, > > I'd be most grateful. > > > > TIA, > > > > > > John > > > > > > > > > > -- > > == > > John Moore - Norwich, UK - [EMAIL PROTECTED] > > == > > ___ > > Rife-users mailing list > > Rife-users@uwyn.com > > http://www.uwyn.com/mailman/listinfo/rife-users > > > > -- > Geert Bevin Uwyn bvba > "Use what you need" Avenue de Scailmont 34 > http://www.uwyn.com 7170 Manage, Belgium > gbevin[remove] at uwyn dot comTel +32 64 84 80 03 > > PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 > Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net > > > ___ > Rife-users mailing list > Rife-users@uwyn.com > http://www.uwyn.com/mailman/listinfo/rife-users > ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] Users guide as PDF?
You can export anything in the wiki (included the live guide) from here: http://rifers.org/wiki/spaces/exportspace.action?key=RIFE On 5-dec-05, at 17:29, John Moore wrote: It would be handy, IMHO, if the users' guide were available as a (preferably printable) PDF. It may well be that Rife provides the facilities to do this already and it's simply a case of Geert pressing the appropriate button? If anyone has generated a PDF from the HTML and could e-mail me it, I'd be most grateful. TIA, John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
RE: [Rife-users] Users guide as PDF?
Hi John, I think you want to have a PDF version of the new 'Live User's guide'. A pdf version of the 'Old User's guide' I have found there: http://rifers.org/wiki/display/RIFE/User's+Guide Cheers, Lars > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of John Moore > Sent: 05 December 2005 17:29 > To: RIFE users list : questions, bug reports and suggestions. > Subject: [Rife-users] Users guide as PDF? > > > It would be handy, IMHO, if the users' guide were available as a > (preferably printable) PDF. It may well be that Rife provides the > facilities to do this already and it's simply a case of Geert > pressing > the appropriate button? > > If anyone has generated a PDF from the HTML and could e-mail > me it, I'd > be most grateful. > > TIA, > > > John > > > > > -- > == > John Moore - Norwich, UK - [EMAIL PROTECTED] > == > ___ > Rife-users mailing list > Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users > ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
[Rife-users] Users guide as PDF?
It would be handy, IMHO, if the users' guide were available as a (preferably printable) PDF. It may well be that Rife provides the facilities to do this already and it's simply a case of Geert pressing the appropriate button? If anyone has generated a PDF from the HTML and could e-mail me it, I'd be most grateful. TIA, John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Hi, [Jython] > I haven't kept up with the recent activity, will they release an new > version soon? I guess not so soon, but there is already Jython 2.2a1. https://sourceforge.net/project/showfiles.php?group_id=12867 Best regards, Lars -- http://semagia.com ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
First of all, an answer as thorough, thoughtful and well-argued as this goes a long way towards convincing me. His response: "I think that's a very precise analysis. As long as people are in denial of a world outside, Struts will seem like the only way. And a rebel framework like RIFE stand as little chance as Rails. But as soon as they open up, I think they want to distance as much as possible (and thus Rails provide the exact opposite of the Struts in a lot of ways)." I can well understand this. I am on the rebound from a not wholly successful project, based around Struts, Spring, Hibernate and various other technologies, where everything ended up taking much longer than anyone envisaged. In the end the company decided to ditch everything and start afresh with Ruby On Rails. I decided not to follow down that path but to stick with Java and part company (I didn't want to waste the benefits of several years of Java experience). Since then I've been assessing various Java based approaches. Those who come from Java are usually tied to the frameworks they already chose and thus very often can't switch to another view on the problem nor another technology stack. If they choose to change, they tend to go for the solutions that already have a large user-base (if 10 people chose it, it can't be bad). Even if they don't get the nicest solution, they at least get one that works well enough and has a pool of knowledge that can be capitalized on. If they choose to change radically, then the memory of the over-zealous complexity of what they already used drives them as far away from Java as they can: RoR. There's probably another factor in there - employability. I am working as a consultant at present, so am able to go with whatever technology seems best for the client. But if I switched to contracting, then Struts/Spring/Whatever skills are far more likely to land me juicy contracts than RIFE skills, at least at present. (This is, of course, a perpetual issue because what is best for one's client is so often NOT the market leader, and often the best paid contract jobs are in areas with complex technologies). Some things have helped people lately: * the URL drives a RIFE application: URL == state * the templates don't drive the pages: they are logicless blueprints When I begin with a new app, I start from the jumpstart and write a simple version of my domain model first. I add RIFE/Crud to get an instant admin interface to be able to easily populate the application with data. Then, either I customize the Crud behavior or write some custom elements. From there on, I extend iteratively and add tests to ensure that the application continues to behaves as intended. If it works for you, it's certainly a good approach for me to start out on! John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Eddy Young wrote: However, as strange as this may sound, I think its Belgian origins also have something to do with its lack of popularity. I think that the IT industry in the US is driven by hype, and since most of it originates from the US, a Belgian web framework does not get to be in the spotlight. This is certainly something I wondered about. 2. How long do you think it would take a reasonably experienced Java developer, who has worked mainly with the Struts/Spring/Hibernate approach to web apps, to get comfortable and productive with RIFE's very different way of thinking/working? One day, if you use all the resources available (wiki, mailing list, IRC) to find the answers quickly. That's encouraging! Thanks for your input. John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
John Moore wrote: I'm encouraged that Jython seems to be showing signs of renewed life (it was falling way behind C Python), as I have really come to like Python for certain kinds of work and the thought of being able to use it in RIFE is appealing. Incidentally, one of the tasks I've been assigned in this new job is to translate some legacy code from Jython to Java. Eddy ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
Not a real howto, but this explains it a bit: http://rifers.org/wiki/display/RIFE/IoC+support http://rifers.org/wiki/display/RIFE/IoC+properties+support+inside +the +repository In this 2d link, the exemple is about a datasource instance created through a property "datasource" using participant "ParticipantDatasources". If I understood, to be able to define a LocaleProvider as a property, with need some participant "ParticipantLocaleProviders". Is it right? No, it could be provided by any participant, and if you add the ParticipantSpringWeb participant, it can come from a Spring IoC setup. However, I'd like to add some basic reference factory behavior to RIFE's IoC, we started design that here: http://rifers.org/wiki/display/RIFE/Add+full+IoC+support You could then do this inside the rep declaration, site structure or element declaration: id="mylocaleprovider">com.uwyn.rife.tools.DefaultLocaleProviderean> mylocaleprovider -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
It's OK, I'm aware of this. By "a complete answer to the questions Ruby on Rails has posed to the Java community", I mean a full-stack framework which has the advantages of comprehensiveness, integration and ease-of-use which people praise in Rails - I wasn't suggesting that the framework itself was written as a response to Rails. One can read a lot of Rails articles suggesting that Java has had its day in the web application area and the future belongs to Ruby, etc. My point was that you can (I hope) mention RIFE to refute this. Thanks, and yes, you can mention RIFE to refute this. Even Rails' creator, DHH, says so: http://rifers.org/testimonials#davidhansson Personally I think that the best thing in RoR is Ruby, I don't like the framework itself very much. Since a wealth of JVM scripting languages are supported in RIFE (like Groovy), you can get much of agility that scripting languages give you. I'm encouraged that Jython seems to be showing signs of renewed life (it was falling way behind C Python), as I have really come to like Python for certain kinds of work and the thought of being able to use it in RIFE is appealing. I haven't kept up with the recent activity, will they release an new version soon? -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Thanks a lot Emmanuel, it's nice to hear how someone got into using RIFE from scratch. Apart from a few extensions you have to make of the Rife classes to take advantage of some Rife features, such as Rife's validation implementations, the framework assumes very little about your code. I still think that this needs to be better and I'll start working on it after the next release (which should be this week, right before JavaPolis), according to the discussion we had about true POJO support (http://thread.gmane.org/gmane.comp.java.rife.user/1245). And if I've not mentioned it yet, the feature I'm loving the most now is Rife's out-of-container testing ( http://rifers.org/wiki/display/RIFE/Out+of+container+testing ); just sweet! :) Yeah, after having added that I had the same feeling. I really feels nice once you start working with it. -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
Geert, either instantiate it at each use. This can then be injected into a property. We still need to add simple reference factory features to RIFE itself currently it delegates that to a repository participant can you give me a link to any howto. Not a real howto, but this explains it a bit: http://rifers.org/wiki/display/RIFE/IoC+support http://rifers.org/wiki/display/RIFE/IoC+properties+support+inside+the +repository In this 2d link, the exemple is about a datasource instance created through a property "datasource" using participant "ParticipantDatasources". If I understood, to be able to define a LocaleProvider as a property, with need some participant "ParticipantLocaleProviders". Is it right? Pierre ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Let me just add that, once you know java, the learning curve is not that steep; it took me about 2 weeks of on/off playing around with the framework to finally "get it". My personal path was: 1. I read and understood the basic building blocks for templates... value(v), block (B, BV), include (I) [ http://rifers.org/wiki/display/RIFE/GuideTemplates ] 2. I followed the in-depth simple blog video, which iteratively introduces most of the key features of rife, essential for app development. [ http://rifers.org/rife_indepth_simple_blog ] After that, I started to move this about, get into the code to understand what's going on for some of the parts (especially with Authentication), asked questions on the list (it's a great community here, and one of the things that got me to get this far into the framework), search the wiki (http://rifers.org/wiki/display/RIFE/Home) and most of it finally makes sense now. Apart from a few extensions you have to make of the Rife classes to take advantage of some Rife features, such as Rife's validation implementations, the framework assumes very little about your code. And if I've not mentioned it yet, the feature I'm loving the most now is Rife's out-of-container testing ( http://rifers.org/wiki/display/RIFE/Out+of+container+testing ); just sweet! :) -- eokyere On 12/5/05, Geert Bevin <[EMAIL PROTECTED]> wrote: > > 1. If RIFE is really all it appears to be (e.g., a complete answer > > to the questions Ruby on Rails has posed to the Java community), > > how come it is not more widely used? It is still a very minor > > player (whether or not this is justified). I was wondering why this > > might be. Could it be to do with its Belgian (i.e., non-US) origins? > > I just wanted to add to this that RIFE never was and still is not > intended to be a RoR-clone for Java. We have been working on the > framework for almost 4 years now and only recently prepped RIFE/Crud > for public release after having used it a long time internally. We > have our own ideals and ideas. Some go intro the same direction as > RoR, but many are also the opposite. It just happens that RoR took > the community by a storm and that people compare RIFE with it. > > Personally I think that the best thing in RoR is Ruby, I don't like > the framework itself very much. Since a wealth of JVM scripting > languages are supported in RIFE (like Groovy), you can get much of > agility that scripting languages give you. > > -- > Geert Bevin Uwyn bvba > "Use what you need" Avenue de Scailmont 34 > http://www.uwyn.com 7170 Manage, Belgium > gbevin[remove] at uwyn dot comTel +32 64 84 80 03 > > PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 > Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net > > > ___ > Rife-users mailing list > Rife-users@uwyn.com > http://www.uwyn.com/mailman/listinfo/rife-users > ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Geert Bevin wrote: 1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? I just wanted to add to this that RIFE never was and still is not intended to be a RoR-clone for Java. We have been working on the framework for almost 4 years now and only recently prepped RIFE/Crud for public release after having used it a long time internally. We have our own ideals and ideas. Some go intro the same direction as RoR, but many are also the opposite. It just happens that RoR took the community by a storm and that people compare RIFE with it. It's OK, I'm aware of this. By "a complete answer to the questions Ruby on Rails has posed to the Java community", I mean a full-stack framework which has the advantages of comprehensiveness, integration and ease-of-use which people praise in Rails - I wasn't suggesting that the framework itself was written as a response to Rails. One can read a lot of Rails articles suggesting that Java has had its day in the web application area and the future belongs to Ruby, etc. My point was that you can (I hope) mention RIFE to refute this. Personally I think that the best thing in RoR is Ruby, I don't like the framework itself very much. Since a wealth of JVM scripting languages are supported in RIFE (like Groovy), you can get much of agility that scripting languages give you. I'm encouraged that Jython seems to be showing signs of renewed life (it was falling way behind C Python), as I have really come to like Python for certain kinds of work and the thought of being able to use it in RIFE is appealing. John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? I just wanted to add to this that RIFE never was and still is not intended to be a RoR-clone for Java. We have been working on the framework for almost 4 years now and only recently prepped RIFE/Crud for public release after having used it a long time internally. We have our own ideals and ideas. Some go intro the same direction as RoR, but many are also the opposite. It just happens that RoR took the community by a storm and that people compare RIFE with it. Personally I think that the best thing in RoR is Ruby, I don't like the framework itself very much. Since a wealth of JVM scripting languages are supported in RIFE (like Groovy), you can get much of agility that scripting languages give you. -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
No, if they are injected through IoC, the reference factory can either provide the LocalProvider as a singleton, singleton: is there no risk of confusion between 2 or more elements from different threads? Not with the interface I proposed: List getLocales(ElementSupport) since each thread will provide its own instance of the element to the method. either instantiate it at each use. This can then be injected into a property. We still need to add simple reference factory features to RIFE itself currently it delegates that to a repository participant can you give me a link to any howto. Not a real howto, but this explains it a bit: http://rifers.org/wiki/display/RIFE/IoC+support http://rifers.org/wiki/display/RIFE/IoC+properties+support+inside+the +repository -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
Geert, Since it should not store the element instance in the LocaleProvider instance. The instance will be constructor through IoC and the element instance is not known in that context. It's also much more appropriate since the element instance changes at each request, and otherwise you'd have to create a new LocaleProvider instance for each request. If we want to avoid creation of LocalProvider instance with each element instance, the only way I see is to store the LocalProvider in a global variable. Is it right? No, if they are injected through IoC, the reference factory can either provide the LocalProvider as a singleton, singleton: is there no risk of confusion between 2 or more elements from different threads? either instantiate it at each use. This can then be injected into a property. We still need to add simple reference factory features to RIFE itself currently it delegates that to a repository participant can you give me a link to any howto. where you either do it in plain Java or interface with an IoC container like Spring or Pico. Regrards Pierre ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] A couple of questions before starting with RIFE
Hi John, 1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? People have said this to me before, but I don't think that's the real reason. Of course any answer to this is pure speculation, but as it turns out I talked about this very issue with Bruce Tate (Beyond Java author) and David Heinemeier Hansson (Rails' creator). This is what Bruce says about it : "I push RIFE every other weekend or so, when I do the nofluffjuststuff symposiums. I also pointed name> toward your frameworks. Usually, the biggest fear is that the project wouldn't get too far without you, so they view it as risky. I'd really work hard toward Spring integration, and work on ways to hide that integration from the user." My remark to David: "I fear though that not many people doing Java are open for the 'quick results without a headache' approach unless they totally switch over to another language or platform. RIFE walking a middle road doesn't help much for its adoption though, but that's another story ;-)" His response: "I think that's a very precise analysis. As long as people are in denial of a world outside, Struts will seem like the only way. And a rebel framework like RIFE stand as little chance as Rails. But as soon as they open up, I think they want to distance as much as possible (and thus Rails provide the exact opposite of the Struts in a lot of ways)." I personally think that RIFE is in a difficult position for both people coming from Java and people coming from PHP. Those who come from Java are usually tied to the frameworks they already chose and thus very often can't switch to another view on the problem nor another technology stack. If they choose to change, they tend to go for the solutions that already have a large user-base (if 10 people chose it, it can't be bad). Even if they don't get the nicest solution, they at least get one that works well enough and has a pool of knowledge that can be capitalized on. If they choose to change radically, then the memory of the over-zealous complexity of what they already used drives them as far away from Java as they can: RoR. Those who come from PHP are used to a scripting language and know of a lot of the monstrous Java-based setups and their complexity. They don't risk getting close to it and go towards RoR instead. A last part that plays against it is that while RIFE aims to provide as much productivity and development fun as possible, we value maintainability and clarity above that. This means that there is a site-structure as a central point of declaration with the data and logic flow inside it. Several people already told me first hand that when they read the features, they loved what they read, but once they saw the site-structure they remembered the XML declarations of other frameworks and concluded that it was just a variation on the same theme, without looking at what is really done in there. 2. How long do you think it would take a reasonably experienced Java developer, who has worked mainly with the Struts/Spring/ Hibernate approach to web apps, to get comfortable and productive with RIFE's very different way of thinking/working? Depends on what you're doing. RIFE is setup (jumpstart) so that you can start right away without having to plan, configure and setup a whole collection or things before you even try something out. So, getting your first results is almost instantaneous and, as Eddy says, within a day you'll have something going. The next stage is a bit more difficult though, since RIFE does a number of things totally differently from what you already know: no servlet-side sessions, logicless templates, declaration of data-flow, url == state, ... I have never been in the situation of having to learn that (since I designed the features), but most people just need time to get the "aha!" feeling. Once they've switched to the mindset we rarely get any questions, except for bugs, holes in the documentation, missing features or complex aspects of the framework. 3. Are there any tips anyone could give me about how to start 'thinking in RIFE'? I.e., how to approach the design of a web application with RIFE in mind. Perhaps something along the lines of planning it as a flow diagram, thinking of possible exits from each page view? Some things have helped people lately: * the URL drives a RIFE application: URL == state * the templates don't drive the pages: they are logicless blueprints When I begin with a new app, I start from the jumpstart and write a simple version of my domain model first. I add RIFE/Crud to get an instant admin interface to be
Re: [Rife-users] A couple of questions before starting with RIFE
John Moore wrote: As I've spent a long time trying to ensure that the set of web technologies I work with next are the right ones, I have a couple of questions about RIFE which I'd like to hear people's responses to. The questions are not meant in any way as a troll, or as an attempt to disparage - I am just expressing genuine questions I have in my head. 1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? It's true that I first heard about RIFE following the Blah Blah list "controversy", but I have never thought of it as a RoR alternative. For me, RIFE is just a web framework, but not one of those that give me the urge to go and write my own. However, as strange as this may sound, I think its Belgian origins also have something to do with its lack of popularity. I think that the IT industry in the US is driven by hype, and since most of it originates from the US, a Belgian web framework does not get to be in the spotlight. 2. How long do you think it would take a reasonably experienced Java developer, who has worked mainly with the Struts/Spring/Hibernate approach to web apps, to get comfortable and productive with RIFE's very different way of thinking/working? One day, if you use all the resources available (wiki, mailing list, IRC) to find the answers quickly. 3. Are there any tips anyone could give me about how to start 'thinking in RIFE'? I.e., how to approach the design of a web application with RIFE in mind. Perhaps something along the lines of planning it as a flow diagram, thinking of possible exits from each page view? Forget about session variables, think in application state. Eddy ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
[Rife-users] A couple of questions before starting with RIFE
As I've spent a long time trying to ensure that the set of web technologies I work with next are the right ones, I have a couple of questions about RIFE which I'd like to hear people's responses to. The questions are not meant in any way as a troll, or as an attempt to disparage - I am just expressing genuine questions I have in my head. 1. If RIFE is really all it appears to be (e.g., a complete answer to the questions Ruby on Rails has posed to the Java community), how come it is not more widely used? It is still a very minor player (whether or not this is justified). I was wondering why this might be. Could it be to do with its Belgian (i.e., non-US) origins? 2. How long do you think it would take a reasonably experienced Java developer, who has worked mainly with the Struts/Spring/Hibernate approach to web apps, to get comfortable and productive with RIFE's very different way of thinking/working? 3. Are there any tips anyone could give me about how to start 'thinking in RIFE'? I.e., how to approach the design of a web application with RIFE in mind. Perhaps something along the lines of planning it as a flow diagram, thinking of possible exits from each page view? Looking forward to your responses! John -- == John Moore - Norwich, UK - [EMAIL PROTECTED] == ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
Since it should not store the element instance in the LocaleProvider instance. The instance will be constructor through IoC and the element instance is not known in that context. It's also much more appropriate since the element instance changes at each request, and otherwise you'd have to create a new LocaleProvider instance for each request. If we want to avoid creation of LocalProvider instance with each element instance, the only way I see is to store the LocalProvider in a global variable. Is it right? No, if they are injected through IoC, the reference factory can either provide the LocalProvider as a singleton, either instantiate it at each use. This can then be injected into a property. We still need to add simple reference factory features to RIFE itself currently it delegates that to a repository participant where you either do it in plain Java or interface with an IoC container like Spring or Pico. -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] localized login template
Geert No, it will be a property, you can set properties inside the tag that will be used throughout the system. You can override them in the site structure or for individual elements. [...] Since it should not store the element instance in the LocaleProvider instance. The instance will be constructor through IoC and the element instance is not known in that context. It's also much more appropriate since the element instance changes at each request, and otherwise you'd have to create a new LocaleProvider instance for each request. If we want to avoid creation of LocalProvider instance with each element instance, the only way I see is to store the LocalProvider in a global variable. Is it right? Pierre ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users
Re: [Rife-users] Tomcat Authentication (was: WebDAV Methods, Sniffing, Authentication)
Fred, what is important for you, to have HTTP Authentiation or to access tomcat's database? I would implement HTTP Authentication using RIFE'd RoleUser credentials and authentication session managers. This gives you container independence and easy migration in case you need it. On 4-dec-05, at 21:07, F Baube wrote: Nope, that's all done in the Authentication element. However, you need nothing of this for HTTP authentication. I don't exactly remember how that works but I suspect that you need to set a couple of headers. It's dead simple; see next paragraph. But the problem now is that Tomcat (version 5.5.12) will not give me access to its MemoryRealm, so I wonder if any of the Tomcat experts on the list can help me out. (See "HOWEVER", below.) "tcpmon" revealed what a Tomcat authentication adapter would need to do. Tomcat replies to an unauthorised request with (for example) HTTP/1.1 401 Unauthorized WWW-Authenticate: Basic realm="Tomcat Manager Application" When the browser sees "401", it ignores the body and puts up a login dialog box. Click OK and the browser begins adding this header to every request: Authorization: Basic ZnJlZDp0b21jYXRkZXJm The browser continues to send this "Authorization:" header with every subsequent GET. The (sample) string "1jYXRkZXJmZnJlZDp0b2" is a base64 encoding (i.e. cleartext) of the string username:password The browser keeps the user logged in until the browser is closed and the browing session ends. So it seems that to emulate this behavior, an adapter would have to check Tomcat's authentication database on _every_ request. ** HOWEVER ... An authentication adapter would ask Tomcat for direct access to the Realm that is in use. But, I find that Tomcat is not being cooperative with its MemoryRealm authentication data :-/ The wiki answers this question: http://wiki.apache.org/tomcat/ HowTo#head-42e95596753a1fa4a4aa396d53010680e3d509b5 Q: How do I get direct access to a Tomcat Realm? A: [..] Note that in order for this to work the Context of the web application in question needs to have its privileged attribute set to "true", otherwise web apps do not have access to the Tomcat classes. [ then sample code is provided ] So, my start-up code makes these calls: Server server = ServerFactory.getServer(); Service svcs[] = server.findServices(); Unfortunately this code is returning a total _zero_ services, apparently/probably because the servlet is not privileged. So, I have tried editing a few candidate files, IAW what I have found on the web and in the book I have. Essentially, I have to mark my servlet's Context ("/mdc") with the attribute: privileged="true" There are three places where this should or could work: * CATALINA_HOME/conf/server.xml * CATALINA_HOME/conf/[enginename]/[hostname] * CATALINA_HOME/conf/context.xml (so that _all_ servlets that are not otherwise configured will be privileged) Unfortunately ... these are ALL failing to grant the privi- lege required ... if indeed privilege is the issue here. Can anyone provide some guidance here ? fred -- F.Baube* Georgetown/MSFS/1988 * Act locally. email fbaube#welho.com * Think pangalactically. gsm +358 41 536 8192 * wmd 60°11'10.8"N 24°57'36.9"E ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users -- Geert Bevin Uwyn bvba "Use what you need" Avenue de Scailmont 34 http://www.uwyn.com 7170 Manage, Belgium gbevin[remove] at uwyn dot comTel +32 64 84 80 03 PGP Fingerprint : 4E21 6399 CD9E A384 6619 719A C8F4 D40D 309F D6A9 Public PGP key : available at servers pgp.mit.edu, wwwkeys.pgp.net ___ Rife-users mailing list Rife-users@uwyn.com http://www.uwyn.com/mailman/listinfo/rife-users