Use offline dictionaries as fallback

Bug #930228 reported by Steve G.
14
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Lingo
New
Undecided
Unassigned

Bug Description

When no internet connection can be established, lingo should use locally installed dictionaries as data source. I use lingo a lot f.e. and sometimes I don't have an internet connection, which makes lingo useless. And I doubt I'm the only one who encounters that problem.

Revision history for this message
Avi Romanoff (aroman) wrote :

Thanks for your report, Steve.

This gets discussed often, and here's what my position is:

Quality is the #1 issue.

The reason I went through the trouble of finding a free online dictionary API, communicating with the organization behind it, and implementing an API client for it is because I don't know of any offline dictionaries with solid quality.

Basically every other dictionary program for Linux uses WordNet, but I and others tend to agree that the results are far inferior from a professional, commercially-driven project that aggregates multiple data sources and spends a great deal of effort assuring quality. (In fact, Wordnik [the API Lingo currently uses] actually uses WordNet as one of its data sources).

Effectively, I see it as a compromise between quality and convenience (or availability).

One potential solution is to, as you suggested, fall back to an offline dictionary when an Internet connection is not available. The obvious problem with this is that Lingo would have different results of different (substantially lesser) quality when offline as opposed to offline.

It's a tricky situation, at best. I'll leave this bug open for discussion, but that's where I stand as of now.

Changed in lingo-dictionary:
status: New → Opinion
status: Opinion → New
Revision history for this message
David Gomes (davidgomes) wrote :

I agree with aroman, we shouldn't have an offline dictionary.

Revision history for this message
Cris Dywan (kalikiana) wrote :

I understand the concern of having different results. That said, I personally do use a separate offline dictionary. It is moot that it may be different if there's no network available. Worse, I tend to prefer the offline dictionary because it works in both cases and I hate having to think about which one I need to use.

Revision history for this message
Steve G. (sgo.ger) wrote :

Thanks for your fast and comprehensive answer, Aroman. I understand the quality concerns, but I strongly agree with Christian here. I tested lingo now for only a couple of weeks, but am about to install another dictionary just because I have to use it offline sometimes.
A dictionary's purpose is to deliver you a definition of a word. Of what use is it if it can't fulfill this purpose under some, common, circumstances? All the nice, quality results it gets are worth nothing if you happen to be unable to connect to the internet.
I know the elementary philosophy is not to try to make it right for everyone, but I don't think that this should include the offline use of the operating system, should it?

Revision history for this message
Andrea Basso (voluntatefaber) wrote :

I think we could use Wiktionary for both non-English languages and as an offline-source, since it's possible to download it.

Revision history for this message
Avi Romanoff (aroman) wrote :

I think Wiktionary has a serious quality issue, and a relatively low signal to noise ratio. I would not be opposed to using it as a source, but only if an API could be written to spit back higher quality results for it. Might be an excellent choice, in that case.

Avi Romanoff (aroman)
Changed in lingo-dictionary:
milestone: none → next
Revision history for this message
Aditya V (kroq-gar78) wrote :

The latest dump of Wiktionary is here: http://dumps.wikimedia.org/enwiktionary/20120320/

The English Wiktionary (pretty sure it's the largest) is about 302MB. I, for one, think that's pretty big.

I haven't done much research (pulled suggestions from here: http://www.linuxquestions.org/questions/linux-software-2/offline-dictionary-like-wordweb-736898/), but some of the offline suggestions were stardict and its stardict-english. Maybe use stardict-english as a dependency and tell Lingo to parse it somehow?

Revision history for this message
Avi Romanoff (aroman) wrote :

Again, the issue here is not quantity or availability of offline resources. It's simply a matter of quality.

If someone can show that stardict or wiktionary or any other offline source offers comparable or better results than Wordnik, then it would be a suitable candidate.

Revision history for this message
Kurt Smolderen (kurt.smolderen) wrote :

I also would like to have offline dictionaries available. What about the following ideas?

1) Would it be a solution to display the estimated quality of a result in some way to the user? I'm currently thinking about using the 'five stars' which are used in music players to rate a song. Lingo sources could in this case get rated and when a result is displayed, the rate given to the source the result is extracted from is displayed to the user

2) Implement multiple sources but distribute them as separate packages. In this case, a user must explicitly install source of 'inferior quality'.

Just some ideas, but I think Lingo really needs to support non-English languages (ref. https://answers.launchpad.net/lingo-dictionary/+faq/1896) and offline usage and this way, you could combine these features in one general design...

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.