Hopbot log for 2008-05-06 - Helma IRC channel: #helma on irc.freenode.net
2008-05-06:
[7:59] <zumbrunn> good morning[7:59] * zumbrunn is in vienna, setting things up
[9:22] <zumbrunn> hi bslivka
[9:22] <zumbrunn> any idea whatelse we could try to get the ichat working?
[9:23] <bslivka> Well I'm trying google
[9:23] <bslivka> for error -22
[9:27] <bslivka> there may be some ports that are blocked
[9:28] <zumbrunn> right, the wlan setup probably stands in the way :-(
[9:28] <zumbrunn> anyway, anything useful, I'll record and make available later
[9:29] <bslivka> excellent
[9:31] <bslivka> rebooting router..
[9:31] <zumbrunn> besides talking about what is currently there and what is planned for helma ng, we'll talk about the following other topics
[9:31] <zumbrunn> (or go into more details on these specific topics)
[9:32] <zumbrunn> how to build libraries in a better way
[9:32] <zumbrunn> what kind of additional libraries to build
[9:33] <zumbrunn> how to make modules better usable for reusable parts of applications
[9:33] <zumbrunn> (vs just libraries)
[9:33] <bslivka> I see
[9:33] <zumbrunn> which additional documentation and tools are needed
[9:34] <zumbrunn> like eclipse integration, or a book project
[9:34] <zumbrunn> (I have a publisher that is interested)
[9:34] <bslivka> I've forwarded some ichat ports, do you mind if I give the ichat one more go?
[9:37] <zumbrunn> ...
[9:37] <bslivka> Oh well
[9:37] <bslivka> sometimes ichat just works
[9:37] <zumbrunn> more generally, how to organize the helma project further
[9:37] <bslivka> Yes
[9:38] <bslivka> perhaps whether to add "ajax support"
[9:38] <bslivka> hehe
[9:38] <zumbrunn> or possibly a separate non-profit that would be a catalyst for the helma project between the different stakeholders
[9:38] <zumbrunn> then about the object/relation mapping
[9:38] <zumbrunn> how to approach that for helma ng
[9:39] <bslivka> Oh I have some things to say about object/relation mapping
[9:39] <bslivka> Some very unpopular things!
[9:39] <bslivka> but opinions nonetheless
[9:39] <zumbrunn> if/how to leverage jython in addition to rhino as part of helma
[9:40] <zumbrunn> yes, I know the object relational stuff is an area where you could contribute ideas
[9:41] <zumbrunn> if something comes to mind that I should put out into the room, just let me know
[9:42] <bslivka> Well I could say the most controversial opinion of mine. It's not particularly constructive, but it may invoke thought
[9:43] <zumbrunn> go ahead
[9:43] <bslivka> ORM is the wrong approach to be taking altogether.
[9:44] <bslivka> That is, as a method of interfacing between a programming environment and a database, it's logically corrupt
[9:46] <bslivka> I may just let that sink in a while, whilst I work on a convincing argument as to why
[9:46] <bslivka> and what I think would be a better approach
[9:47] <bslivka> it's a big subject, may be out of scope
[9:48] <zumbrunn> I've read some blog posts from you that make me understand what you mean in general
[9:51] <bslivka> Really most of my opinions about ORM come from chris date, mostly because I think he's made a convincing argument
[9:52] <bslivka> But the place where it really hurts helma is that helma's abstraction doesn't quite go far enough, and exposes SQL to the programming environment via the type.properties
[9:53] <bslivka> and the differences between the xml database and the SQL database mappings are symptoms of this deeper problem
[9:55] <bslivka> the way I see it, in javascript we have String, Array, Object, Number, Boolean and Function
[9:56] <bslivka> Number is represented in the implementation as a IEEE float
[9:57] <bslivka> Logically, to the program, the idea is to just treat it as a platonic Number, with typically no real limitations
[9:58] <bslivka> the implementation IEEE float has limitations of course, but the programmer is not asked to understand in detail what they are
[9:59] <bslivka> The trick, is to produce two types, something akin to Number
[10:00] <bslivka> but simply called Relation and Tuple
[10:00] <bslivka> which serve as logical mathematical abstractions to the underlying implementation, in the same way that Number is an abstraction of float
[10:02] <bslivka> on the surface it sounds kind of the same as Hopobject and collection, but there are numerous little nuances that make it different.
[10:03] <bslivka> and to be perfectly honest I haven't worked out in detail exactly how it can work in javascript.
[10:04] <bslivka> but the basic message is that having something that represents a mathematical relation, complete with a relational algebra that doesn't involve SQL is something that I'd like to see someday
[10:04] <bslivka> whether from someone else, or when I figure out how to do it myself.
[10:05] <bslivka> and it's looking like with the work that Maks is doing on helma, in combination with the changes in Helma NG, my ideas are looking more and more easy to implement
[10:05] <bslivka> given that Helma NG looks like it will allow javascript implementations of persistance routines.
[13:32] <mmonster_mac> meetup still ongoing?
[13:58] <zumbrunn> yep it is
[13:58] <zumbrunn> hi mmonster_mac
[13:58] * zumbrunn is in the middle of a discussion right now
[13:58] <zumbrunn> just a sec
[20:02] <midnightmonster> *sigh* I've got an e4x bug in the rhino used in 1.6.2
[20:03] <midnightmonster> sometimes (haven't pinned down when) it will put quotes around values included in tag content as though it were including the value in an attribute
[21:33] <bretonslivka> mornin
[21:37] <bretonslivka> meet go okay?