[Lift] Re: Is Lift the right tool for us?
On Tue, Apr 21, 2009 at 7:12 AM, Jeppe Nejsum Madsen je...@ingolfs.dkwrote: Hello, I've been lurking on this list for a while and trying out small things with Lift/Scala but thought I would ask the more experienced people around here before digging much deeper. First some background: I've been writing webapps for well over a decade using all kind of tools/languages ranging from Perl/PHP to ASP/.Net/Java (but never Python or Ruby in any significant way). The last 5 years I've mostly focused on Java/JEE. Over time I've, like many of you probably, grown tired of the verbose java syntax and the xml bloat that seems to be required to get any kind of real-life system up and running. I'll soon be starting on a new project where I can choose the entire technology stack and I'm currently evaulating the options. The app will not be a new Web 2.0 social platform used by millions, but rather a more boring business-type app with lots of data entry and interactive reporting :-) But it will have to stable, fast, easy to use (ie will probably need lots of Ajax functionality) as well as integrate with many external parties. One of the primary parameters for choosing a language/framework is speed of development. I know Rails excel in this and may be the sweet spot for an app like ours, but 1) I don't know ruby that well and 2) I've tried RoR and it never really clicked with me while I can easily see how you can create something, fast. Python/Django is another option I'll be looking at. I don't know Lift/Scala either but I've always liked the clean, simple syntax of SML (did a fair share of SML programming back in the CS classes :-), so FP is not uncommon to me (but a bit rusty). And I do like the JVM as platform. So some of the questions I have to the community is this: - How long time does it take for experienced developers to be 1) Productive and 2) Be hyperproducive :-) if we assume we'll start from scratch? 1 - About a month 2 - About 3 months - How about new members to a team already knowledgable about Lift/Scala? It's not likely we'll be able to hire lots of experienced Lift developers in the near future :-) The same as for RoR or DJango. Your curve will be similar. New language, new framework. The plus is that you don't have to worry about operations because Lift deploys as a WAR file. Learning Capistrano for Rails is non-trivial. - How is the productivity compared to say Rails Django for database driven apps requiring lots of Ajax? Better. Lift's Ajax and JavaScript abstractions are worlds better that Rails and Django. Lift abstracts away the HTTP request/response cycle. Rails puts you in the heart of it. With Lift, you need not define 3 separate files to handle a single Ajax request. - Is Lift ready for production use? Yes. Lift apps are in productions (and have been for 1 year) at companies including SAP, Enthiosys, and other places. More Lift apps are coming online every day and companies that you've heard of (e.g., Siemens). While our project is not mission critical requiring 5+ nines, it is not a consumer product either. Downtime will hurt. I've read David's statement that he will do a lot to support production installations, which is a very nice commitment. I know this depends on the app architecture, but would be nice to get some - Any input on Mapper/Record vs JPA for persistence? Depends on your app. Thanks, David But basically any experience from using Lift on real, substantial projects is welcome. /Jeppe -- Lift, the simply functional web framework http://liftweb.net Beginning Scala http://www.apress.com/book/view/1430219890 Follow me: http://twitter.com/dpp Git some: http://github.com/dpp --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
Thanks for the swift reponse. A few comments inline On Tue, Apr 21, 2009 at 7:51 PM, David Pollak feeder.of.the.be...@gmail.com wrote: [...] I don't know Lift/Scala either but I've always liked the clean, simple syntax of SML (did a fair share of SML programming back in the CS classes :-), so FP is not uncommon to me (but a bit rusty). And I do like the JVM as platform. So some of the questions I have to the community is this: - How long time does it take for experienced developers to be 1) Productive and 2) Be hyperproducive :-) if we assume we'll start from scratch? 1 - About a month 2 - About 3 months This is encouraging! - How about new members to a team already knowledgable about Lift/Scala? It's not likely we'll be able to hire lots of experienced Lift developers in the near future :-) The same as for RoR or DJango. Your curve will be similar. New language, new framework. The plus is that you don't have to worry about operations because Lift deploys as a WAR file. Learning Capistrano for Rails is non-trivial. - How is the productivity compared to say Rails Django for database driven apps requiring lots of Ajax? Better. Lift's Ajax and JavaScript abstractions are worlds better that Rails and Django. Lift abstracts away the HTTP request/response cycle. Rails puts you in the heart of it. With Lift, you need not define 3 separate files to handle a single Ajax request. Yes, the Rails way seem a bit cumbersome (but does provide a quick bang for the buck out the box). I've yet to fully understand how this works in Lift (Scala code still doesn't look very clean to me. Maybe that's just an acquired taste?) [...] - Any input on Mapper/Record vs JPA for persistence? Depends on your app. Ok, will get back to this when I've got some more details... /Jeppe --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
On Tue, Apr 21, 2009 at 9:01 PM, Charles F. Munat c...@munat.com wrote: Yeah, it's always nice to get some... :-) As a former Rails developer, I'll say this: I don't think I'm yet quite as fast in Lift as I was in Rails, but I'm getting there. A lot of this has had to do with switching mindset from Ruby to a more Java-ish approach. Do you think you'll surpass your Rails speed? If yes, can you explain what elements contribute to this? For you, as a Java developer, I think you'll find the reverse true. With Scala, you can continue to use the same Java libraries and tools that you've used before. I think you'll learn Lift very quickly, and whenever you're stuck, you'll have all that Java background to fall back on. My thoughts exactly :-) I know I will not use Java for this project so I'm willing to take a small hit initially. But staying on the JVM at least provides som firm ground wrt to deployment, ops, runtime, tuning etc. If your project is more than a simple app (as I'm certain it is), I would go with JPA. Record is not ready yet, and Mapper is pretty limited. For example, AFAIK Mapper doesn't do many-to-many associations. I've had a lot of success with JPA/Hibernate, and it's getting pretty solid in Lift now -- there are even Maven archetypes for it. I've yet to dig into Mapper/Record. What are the main benefits of these (besides the obvious simplicity you get when not needing to use JPA)? Lift is more than ready for production use. I've been using it in production for more than six months. Nice! Any numbers you can share (users, tables, data, transactions etc)? /Jeppe --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
If your project is more than a simple app (as I'm certain it is), I would go with JPA. Record is not ready yet, and Mapper is pretty limited. For example, AFAIK Mapper doesn't do many-to-many associations. I've had a lot of success with JPA/Hibernate, and it's getting pretty solid in Lift now -- there are even Maven archetypes for it. I've yet to dig into Mapper/Record. What are the main benefits of these (besides the obvious simplicity you get when not needing to use JPA)? Here's my experience: I started out with mapper for my current project, and ended up overriding/rewriting most of the functionality, to make it work for my (maybe unusual) requirements. I have pretty much duplicated the entire mapper class/trait hierarchy, bit by bit. For example, Mapper does not like dealing with null values for types that are not nullable in Scala. Mapper has, for example, 0 hard-coded as default value for int's for which no value is given. No insurmountable obstacles, but things might take more work than one thinks initially. There's just no free lunch. Since mapper combines so many aspects, it might take a considerable amount of effort to isolate concerns when switching from mapper to JPA or similar. I will be switching to JPA in the next two weeks. Best, Clemens --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
I too used to do a lot of rails dev... Granted my lift apps take me *slightly* longer to produce, but only by a bit. However, they do take a lot less time to deploy and maintain because my platform is not leaking like the proverbial rusty bucket ;-) Cheers, Tim On 21/04/2009 20:15, Jeppe Nejsum Madsen je...@ingolfs.dk wrote: As a former Rails developer, I'll say this: I don't think I'm yet quite as fast in Lift as I was in Rails, but I'm getting there. A lot of this has had to do with switching mindset from Ruby to a more Java-ish approach. Do you think you'll surpass your Rails speed? If yes, can you explain what elements contribute to this? --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
Jeppe Nejsum Madsen wrote: Do you think you'll surpass your Rails speed? If yes, can you explain what elements contribute to this? I'm not sure, but I think so. Then again, it's difficult to say, because I'm doing more. There is definitely a bit less boilerplate for me with Lift, but not as much less as I'd like. One of the things I'm slowly working on is some best practices to reduce that boilerplate. But I build apps very differently in Lift than I did in Rails. Not just because Lift is different. Mostly, it's because I've refined my ideas and new libraries have made things easier. Where Lift sometimes slows me down is usually because I'm learning a Java library. For example, JAI has wasted a lot of my time to do something that I did fairly quickly with RMagick in Rails. But I can do a lot more with JAI once I have it down. That's the tradeoff, I guess. In Rails I wrote an extensive library of CRUD functionality that I used to build CRUD apps very quickly. My views were reduced to a few lines of code, controllers even more so. For example, a controller that had 350 lines of code was reduced to around 12 lines. A view that had 200 lines became 20 lines. That really made things fast. I hope to do the same for Lift eventually (if someone else hasn't done it first), but in the meantime, I'm not quite as fast. For you, as a Java developer, I think you'll find the reverse true. With Scala, you can continue to use the same Java libraries and tools that you've used before. I think you'll learn Lift very quickly, and whenever you're stuck, you'll have all that Java background to fall back on. My thoughts exactly :-) I know I will not use Java for this project so I'm willing to take a small hit initially. But staying on the JVM at least provides som firm ground wrt to deployment, ops, runtime, tuning etc. Oh, you'd be surprised how handy having access to Java at any time is. You get all the benefits of Scala's syntax, but you can use pretty much anything out there that's Java -- without writing any Java. That's probably what persuaded me to try Lift more than anything else. I hate Java's verbosity (and all those damn XML files). But I love its speed and stability and the huge amount of code out there. If your project is more than a simple app (as I'm certain it is), I would go with JPA. Record is not ready yet, and Mapper is pretty limited. For example, AFAIK Mapper doesn't do many-to-many associations. I've had a lot of success with JPA/Hibernate, and it's getting pretty solid in Lift now -- there are even Maven archetypes for it. I've yet to dig into Mapper/Record. What are the main benefits of these (besides the obvious simplicity you get when not needing to use JPA)? I don't find Mapper any simpler than JPA, just different. You do get some stuff for free, but it doesn't ever seem to be what I need. Record is another matter. I haven't even looked at it yet, but from what I hear it bundles in oodles of goodness. I'm psyched. And I think you'll be able to use it with JPA... could be wrong about that, though. Lift is more than ready for production use. I've been using it in production for more than six months. Nice! Any numbers you can share (users, tables, data, transactions etc)? Funny thing: I have no idea. I build sites for clients and I've never concerned myself with traffic. I don't think any of them are very high traffic, though. Maybe a few thousand hits a day. Maybe more. But David Pollak has done some amazing stuff, so check with him on that. Chas. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
[Lift] Re: Is Lift the right tool for us?
On Tue, Apr 21, 2009 at 1:53 PM, Charles F. Munat c...@munat.com wrote: In Rails I wrote an extensive library of CRUD functionality that I used to build CRUD apps very quickly. My views were reduced to a few lines of code, controllers even more so. For example, a controller that had 350 lines of code was reduced to around 12 lines. A view that had 200 lines became 20 lines. That really made things fast. I hope to do the same for Lift eventually (if someone else hasn't done it first), but in the meantime, I'm not quite as fast. This kind of support is one of our goals with Record. There's still some mismatch between the way that JPA sees the world and the way that Record sees the world, but I'm confident that we can find some way to get them playing together nicely. Derek --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Lift group. To post to this group, send email to liftweb@googlegroups.com To unsubscribe from this group, send email to liftweb+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~--~~~~--~~--~--~---
