Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Menu
Murphy law: it's when you are writting a long wall post about the new version, all it's feature, current limitation etc. that your firefox will crash, and in a way it has lost the message
So well before I get back the motivation to rewrite it:
here is the link
http://tato.sysko.fr/ , it's quite featureless (I will detail them after), so waiting my post, don't lose time saying "this or that is missing" :)
For the moment it's mainly to show you the improvements
briefly
speed improvement
realtime updated searchengine
translations showed up to a depth of 20 (or 5 on some pages) (but it's me who limited it, no more due to performance/architectural limitation)
realtime duplicate detection/merging
the database is an old old one, the user accounts has not been imported so feel free to register new accounts
it's on my own server, not the same as tatoeba, so feel free to torture it and report problems
PS: the selects doesn't show all the languages supported because I've still a nasty bug to fix with the search engine
Cool!
Also to you i recommend https://addons.mozilla.org/de/f...form-recovery/
* Hope for the problem of duplicates :-) *
The new version http://tato.sysko.fr/eng is just WONDERFUL!!! Thank you very much, sysko! Certainly much more than a lot of work...
So now, if someone does not want to create a duplicate and wants to know, if a certain sentence has already a translation in a given language - easy game, he has just to go to http://tato.sysko.fr/eng/sentences/show/37770 (with the right number in the end). We will have to ignore the mention "you're not supposed to view this", but we all ignored lots of similar advices in our lives, no problem :-)
On the page it says, we should "Please repport any bugs/suggestions to 'team @t tatoeba dOt org' ". But sysko stated above: "it's quite featureless (I will detail them after), so waiting my post, don't lose time saying "this or that is missing" :)"
Be aware that the database is quite old. So there may be newer translations in the translation graph.
I found two graphs with a maximum distance of eight, the above mentioned http://tato.sysko.fr/eng/sentences/show/37770 (five English versions - or something similar to English, three in Esperanto or something similar...) and http://tato.sysko.fr/eng/sentences/show/685711 . Are there higher distances in the corpus?
actually the "you're not supposed to view this" is for me, as I use a template system, so if I view this it's to remind me to replace it by the actual right side modules :)
I've started listed all the missing stuff here
http://redmine.sysko.fr/project...1&tracker_id=2
In the future (and even now) I will use it for managing feature request, bugs etc. If I've configurated correctly, it should be open to everyone, so feel free to add tickets (I will try to organize them later). I will soon list all the tickets which are part of the "first release" of this new version.
For the moment I'm trying to organize my new free time between tatoeba (the new version, discussion on the IRC channel, taking care of the current database) and my real life activities. Hope to have a stable schedule soon :)
I will try to replace the version of the database by a more recent one, so that it will be more useful for those who use it to view linking possibilities.
For the maximun distance, here is one that reach 20
http://tato.sysko.fr/eng/sentences/show/60
Actually it's certainly more than 20, but for the moment I've fixed to 20 the maximun distance displayed, I will code after something to permit those who wants to view more
It would be cool to see the translations tree-like, so one can follow where theres a "translation bug" (or an ambigouity) if somewhere the meaning totally changes.
Maybe there could be a button to expand/collapse one more/less hirachy.
Actually the first version I've made was a tree-like, but as a default view, it's not convenient, but yep I can see to add this as an alternate view.
It should be relatively easy to have a toggle button and using jquery to switch between the two. In the end its just:
1, 2, 3, 4, 1, 2, 3, 4 <=> 1, 1, 2, 2, 3, 3, 4, 4
right? If the sentences had a different HTML-class than "directTranslation" like "translation_1" etc. the class could be used to rearange the sentences (for now the number could be used, but i dont know if the number will be kept). badly we would need a fallback for non-JS users.
In the case the starting structure would have been the tree, yep, but for some space optimisation reason, I don't keep this information when getting the sentences from the database. As anyway the informating bringing by the tree are not usefull for 95% of users
So I think i will adapt the following solution
1 - by default it display the "by distance" grouping
there will have a toggle button
the first time it will retrieve the tree structure of that group of sentences, (which is a little more heavy request than the "grouping by distance" one) and it will display it,
and after if you toggle it will be able as you said to do that only on client side, as we will have enough information to switch from one to an other
for people without JS the toggle button will simply redirect to a "show-as-tree" (or any fancy name) page with the tree structure,
the link will be simply overridden by the JS.
that sounds good. i didnt think of the request load, because i dont know too much about databases or their performance. Im just a user, craving for more features ;)
don't worry, that's the deal, you torture your sould to find new features and I torture mine to code it to statisfy your endless hungryness for fancy stuff :P
and the expand/collapse would do this:
1, 2, 3, 4, 1, 2, 3, 4 <=> 1, 2, 3, 1, 2, 3
or this
1, 1, 2, 2, 3, 3, 4, 4 <=> 1, 1, 2, 2, 3, 3
respectively.
Thank you for the link with a long distance; the low number of your sentence gave me the idea to search in the beginning.
http://tato.sysko.fr/eng/sentences/show/3 has at least a maximum distance of 20,
http://tato.sysko.fr/eng/sentences/show/4 as well and http://tato.sysko.fr/eng/sentences/show/28 ...
Really funny to see how these Chinese_whispers (Stille Post, téléphone arabe) change the sentences :-)
By the way: It seems the graph http://tato.sysko.fr/eng/sentences/show/28 has about one thousand sentences... Some languages must be dozens of times in it...
I thought about it, but now I rejected the idea of drawing a graph with one thousand sentences :-)
As the code to manage the database can run without the website, I will try to propose some code example of tools manipulating the database (doing one that output the graph of the group of a sentence in "dot"[1] format that can be after converted into an image, should not be that complicated )
I've updated the database used on the new version, now it is only an hour old :)
this way it should be more usefull for those want to use it to help you link
now I think about it, I can try to keep it weekly updated (the script to feed the graph database use the weekly export anyway) and if someone feels ready to make a little user js script that replace for those who wants the link to ... link on tato.sysko.fr by the corresponding one on tatoeba.org, maybe that could ease a lot those who wants to help creating links.
What do you think ?
Merci beaucoup que tu as mis à l'heure la base de donnée!
Excuse-moi, s.t.p., sysko, en lisant ton commentaire ci-dessus j'ai l'impression que ma capacité de comprendre l'anglais est devenue plus faible ces derniers jours ;-)
Est-ce qu'il serait beaucoup de travail de proposer un lien à ton http://tato.sysko.fr en dessous de chaque liste de traduction? Donc par exemple en bas de
http://tatoeba.org/eng/sentences/show/426660 (4 traductions) on trouverait le lien
http://tato.sysko.fr/eng/sentences/show/426660 (9 traductions).
I'm reconfirming that the duplicate detector seems to work, at least in English. Also, Sysko, are you going to blow off this version of the database when you migrate TatoebaCPP to this domain?
sysko: "translations showed up to a depth of 20 (or 5 on some pages) (but it's me who limited it, no more due to performance/architectural limitation)"
So it would be able to show even more translations. Would it be possible to put a button to see more than 20 translations?
By the way: Does it make sense to always show 20 translations? Wouldn't it be better - consume less power - just to show a depth of e.g. 5 plus a button to see more? I doubt that a lot of people will read one thousand translations...
(And a very small bug on this wall: There is a triangle on the right hand side below to change the size of the window; when you put it to the right, you can't see the triangle any more.)
Yep it will be possible to view everything :) with a show more button.
For the moment I always show 20 translations because it's mostly used by advanced users and I haven't implemented a view more button. Once it will be implemented I will turn the default distance to a lower one.