Eana Eltu: Translator, Dictionary, API and putxìng.

Started by Tuiq, January 07, 2010, 04:20:17 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

`Eylan Ayfalulukanä


Yawey ngahu!
pamrel si ro [email protected]

Tuiq

I just don't know. Of course there would be hosters that deal specifically with this kind of thing, starting from $6.50/month upwards. They would - likely - allow me to simply develop on my local machine, upload it and it would just work. No messing around with Mono, Linux or all that stuff...

However, my past experience with such shared hosters has not exactly been great. I might be able to pull a few strings over a friend of mine, but so far it looks rather dark. It's a pity, because I really do like C# and I think this could be rather nice.
Eana Eltu: PDF/TSV/jMemorize

`Eylan Ayfalulukanä

I am assuming you mean 'Windows server hosting' here. I know little about Windows hosting, as I generally use Linux hosting and avoid Windows at all costs.

I used to program a lot back in the very early days of small computers. Its just one of those things that dropped by the wayside due to lack of time.

Yawey ngahu!
pamrel si ro [email protected]

Tuiq

Yeah, it's the Windows part that is worrying me. I would have several friends with Linux where I could probably come under and of course LN.org's own BSD - but sadly, .NET on those is kind of... err.

Debian, as example, has a quite "outdated" Mono. I would like to develop it with ASP.NET MVC 4 and .NET 4.5 - the best Mono in Wheezy can offer is MVC2, there isn't any package that would offer the recent 3.2 Mono port, which might be able to handle it. Hell, if necessary I might be able to survive with 4.0 or even MVC3, although I really like some of the features in those newer versions.

I'm not a Linux guy at all, I don't know a lot about how binaries or software on Linux works, dependencies and all that - all I see, more or less, is that any version that can be installed without breaking seven warranty seals is three to four years old.
Eana Eltu: PDF/TSV/jMemorize

baritone

Quote from: Tuiq on February 03, 2014, 04:16:02 PMDebian, as example, has a quite "outdated" Mono.
All packages of Debian are slightly outdated, not only Mono.
In order to use newest packages, one ought to use another distributive.
I like "Calculate" - easy to install Gentoo based distributive. You can install Mono 3.2 on it.

Tuiq

To be honest, I'm way past the event horizon with Mono. I'll settle for a Windows server, they're not that expensive I suppose.

Right, so that means also back to the drawing board. While the old new system did kind of okay, it also had some (severe) flaws. The database design wasn't entirely what I wanted it to be - it had recursions in it (if I remember properly) which made updating/erasing data sets a pain.

So that needs to be fixed somehow. I'll review what I made so far and then draw up some intelligent conclusions or something. If anybody has anything to say about the previous attempt/what they expect and I didn't mention here/has questions in general about EE.NET, now's your time to shine!
Eana Eltu: PDF/TSV/jMemorize

Tuiq

A quick mentioning that I do have a capable server now so the project could be executed, in theory, and even made accessible for beta testing and all that stuff.

However, I would start from scratch (as always, the things you saw here were pretty little demos that got me into the material - and I'm still not done with the first tutorial yet, so much to learn!). Not just programming but also design-wise. There are some things that I simply do not agree with.


  • Get rid of the idea that it should be a table. Use panels instead. Maybe each of those panels could be formatted as you please, or simply have fields "highlighted". I have an idea in my head, not seen it physically yet however so I am unsure how well it would work.
  • The whole dictionary-inheriting stuff is nice and all, but I'm not too certain on it either. Mainly the issue that root dictionaries do completely break this cycle is a bit... annoying. There's not too many cases where I could see this being useful; however the current "root dictionary / translation" relationship is not ideal either.
  • The question of interest, of course. I'm out of the loop badly, you could say. I read parts of these forums from time to time, but I don't have stats (absolutely a TODO) on the community, the dictionaries (both Na'vi and Dothraki, but I think GoT is pretty popular). I understand that the current system both works and is a pain in the arse - something that should be changed. But it should be done properly, which for the current code base isn't possible. If you think about how it all began, it's astounding it's still around... five? years later.
Eana Eltu: PDF/TSV/jMemorize

`Eylan Ayfalulukanä

Ma Tuiq, I am not sure what you mean by 'dictionary inheriting', but if it is the ability to build other language dictionaries from the root dictionary, that is perhaps the single biggest thing that separates EE from something you could do in a WYSIWYG word processor, or more traditional LaTeX textbuilding.

Your 'panel' idea is intriguing, especially if it lets one build definitions with just the fields that the definition requires. If possible, you might want to build a sample, so that all the EE users can see what you have in mind. (I am assuming though, there is still a database at the root of everything!)

This is a great time to work on this, as both Na'vi and Dothraki are quiet.

Thanks for not forgetting about EE!

Yawey ngahu!
pamrel si ro [email protected]

Tuiq

I try to not repeat myself with programming speak and instead explain it short and more carefully:

"Inheritance" simply means which field is picked for the final output. Right now, it's a simple parent-child inheritance: If a child overwrites a field (usually the translated one), in the child dictionary that field has priority. It's "overwriting" the parent's (i.e. the root dictionary, the English) field.

The idea which I'm uncertain of is if we could chain this: For example, [enEN] < [frFR] < [chFR]: A field defined in chFR will overwrite any field written in frFR or enEN, while a field written in frFR will overwrite any field in enEN.

"Field" is currently anything between "{}", such the translation, the source, the commentary, anything you can input. Every text box is a field.




Panel: It wouldn't change how the data is handled. The idea is still that you have a set of pre-defined templates (I call them "word types", "templates", such as "word", "wordN", "affixMarkerNNNN902310").

So, how can I break this down... I probably can't. I'll have to show it to you, although the screens I've posted so far should be still somewhat accurate, I think I can improve the design a bit.

I have a few projects queued up first and not too much time on my hands, but it's still something that I wish to do and now that I have the possibility to realise it, it's probably going to happen.
Eana Eltu: PDF/TSV/jMemorize