Greetings, new guy coming through. I am breaking a huge taboo by not having lurked long enough before contributing to fully understand the scope of what is being discussed, but that won't stop me from making an ass of myself first time through! So, here we go.
Having had the good fortune of working with j2ee a bit too long for my taste now, and having had the good fortune of being in the VP and CTO role, as well as having returned to the trenches for coding, I definetly have an opinion about this... though I don't know if it even applies. I personally believe that p5ee will *not* be taken seriously without a series of core requirements being fulfilled in the specification (at first, and in the code second.) If in the _artifacts_ it is stated that the desired target platform for p5ee is a Perl that has functioning unicode support, threading (i see less need for threading, but that's because i happen to think it's a piss poor mechanism for parallelization of tasks, but that's another discussion on some other mailing list), component design patterns, component managed serialization, etc... (all of those things those javasoft bastards suport :), it is a step in the right direction. Now, notice I said that it didn't *have* to be supported at first, just that the design decisions were based on a platform that provided this minimum support and functionality. This doesn't mean that valuable code can't be generated for earlier, less "feature wealthy" platforms, as a matter of fact, the effort *will* have to be on those platforms, as our goal platform doesn't even really exist yet... but I think that the intent should show that P5ee *will* have all of the basic requirements for being "enterprise grade". I mean... how many features are some of the other "enterprise vendors" supporting that are currently not even implemented in any container or supporting platform? Their requirements and specification currently outstrips the product offering... *currently*... they've shot higher than technology can support *right now*... and I think that's the right way to do it. Shoot as high as you can, and then catch up to your goal. :) Just my greatly uninformed 2 cents. Eduardo > Since perl is free to download and install, I don't see a big issue with > requiring the newest version. They will have to install the p5ee packages in any > case. > > There are some disadvantages to allowing 5.005_x or older: > > 1. Older than that and you get back to the time when the windows and unix > versions diverged greatly. This could seriously impact our ability to serve the > win32 group > 2. If we allow older versions of perl, we'll need to spend a lot of time making > sure our code works with all those older versions of modules that came with the > base install. > 3. Some critical things for the enterprise, such as unicode and threading are > only just now beginning to stabilize. I can't see how we can claim our perl EE > is truly enterprise ready without those key elements. > > We could always create a mini perl install which contains the required version of > perl and all our modules. That way, all the new code would work correctly, and > all the legacy stuff would continue to operate. > > Stephen Adkins wrote: > > > At 03:05 PM 1/2/2002 +0000, Matt Sergeant wrote: > > >On Sun, 20 Jan 2002, Gerald Richter wrote: > > >> Until now I thought the consensus on the list was to also support 5.005 ? > > >> > > >> Since this is the second time we have the discussion whenever to use > > >> attributes or not and therefore to require Perl 5.6 I think we should > > >> clarify this. Vote on it? > > > > > >I vote we support 5.8+ only. Seriously. > > > > Hi, > > > > This is the second vote called. > > We'll do better than last time, hopefully (in terms of process). > > Basically, there are two phases to this: > > > > 1. the discussion to formulate the questions for a vote > > 2. the vote > > > > We are in the discussion phase now. > > Here are some questions I think we need feedback on. > > > > QUESTIONS: > > * What benefits would result from raising the required level to 5.6.1? > > (Someone already commented that the bugginess of 5.6.0 would preclude > > us using that as a supported base.) i.e. What modules would we desire > > to use that we could not if we had to support 5.5.3? > > * Same question for 5.8.0? > > * What about mod_perl? Are 5.8.0 and 5.6.1 both as well supported in > > mod_perl as 5.5.3? > > * What versions of Perl ship with major Unix systems? Linux distributions? > > * How stable is 5.8.0? > > > > PRELIMINARY OPINIONS: > > I think the success of P5EE is not dependent on features beyond 5.5.3. > > I suggest that P5EE 1.0 should be based on a pervasively deployed 5.5.3. > > If we can get a body of code out there that functions and begins to gain > > acceptance, we might consider P5EE 2.0 to be built on 5.8.x. > > > > The outcome of any meaningful vote will result in a modification to the > > artifacts (documents/code) of the project. > > > > Part of this discussion phase will include me floating the questions > > for a vote, and I need feedback on that along with discussion about it. > > > > I think the questions are the following (which would be put to > > two separate votes of +1, +0, -0, and -1). > > > > 1. Do we change the Perl Version supported by P5EE from 5.5.3 (5.005_03) > > to 5.6.1? > > > > 2. If we vote to change the Perl Version supported from 5.5.3, do we > > change it to 5.8.0? > > > > These decisions affect assertions made in the following documents. > > > > In the P5EE Perl Style Guide, we make our current statement that 5.5.3 is the > > supported Perl Version > > > > http://www.officevision.com/pub/p5ee/software/htdocs/P5EEx/Blue/perlstyle.ht > > ml#perl%20version > > > > This Perl Version was selected previously in order to maximize portability. > > "Pervasive Deployment" is one of the core elements of the vision of P5EE. > > > > http://www.officevision.com/pub/p5ee/ > > > > Also, Portability is one of the Attributes of Enterprise Systems. > > > > http://www.officevision.com/pub/p5ee/definitions.html > > > > Furthermore, the ability to run in mod_perl is one of the key requirements > > based on expected platforms. > > > > http://www.officevision.com/pub/p5ee/platform.html > > > > Stephen
