Which Sentences I Use and Which I Do Not Use

I Use These ("Whitelisted")

** A Select Group of Sentences from the #01 WHITE LIST

• audio - eng - by CK - CC BY-NC-ND 3.0 - www.manythings.org/tatoeba.
  This is a list of all the sentences for which I have recorded audio.
  If you are planning only to use a smaller subset of the ! #01 WHITE LIST mentioned below, then this would be the set to use.

1. Sentences I have chosen to use.

Others developers who are using English sentences from the Tatoeba Project may wish to start with this list as your default list of sentences to use. You can grab all the sentences numbers for list 907 from this file (sentences_in_lists.tar.bz2) which is exported every week a little after 9:00 GMT.

• ! #01 WHITE LIST - Proofread Good English Sentences That CK Uses on His Projects - #907
• collections/of/CK/ok - This is not a list but is TRANG's new rating system. It is mostly just a duplicate of List 907.
• (Perhaps there are times when these two are not exactly the same.)

I Do Not Use These ("Blacklisted")

In addition to sentences on these sorted lists, I do not use any sentence that is not on my WHITE LIST (list 907).

2. Sentences I have chosen not to use, but possibly will use in the future for another project.

These should probably all be proofread again by any developer who wants to use these.

• ! #04 FILTER OUT for now - GRAYLIST - maybe use these in the future - good sentences with non-wildcard names - wildcard version may or may not exist - #3642
• ! #05 FILTER OUT - near duplicate - WILDCARD version exists - see bit.ly/tatoebawildcards - #4005
• ! #06 FILTER OUT - near duplicate - CONTRACTIONS preferred - #1159
• ! #07 FILTER OUT - near duplicate - NO EXCLAMATION MARK (!) is preferred for these out-of-context examples - #1237
• ! #08 FILTER OUT - non-American spelling, vocabulary or usage - possibly OK in other dialects - #1394
• FILTER OUT - similar to something already on the WHITE LIST - #4006

3. Sentences I have chosen not to use, at least for now, but may want to review again in the future.

I'm unlikely to actually review these again, since it's a more efficient use of my time to just contribute good English sentences. However, perhaps other developers might want to look at these sentences. I put sentences on various lists to help save you time, since you can prioritize the order in which you want to review these.

• ! #02 FILTER OUT - unlikely to use - possibly OK or almost OK - #3122
• FILTER OUT - near duplicate - better exist, only pronoun difference, or other slight difference - #3012
• FILTER OUT - near duplicate - he/she where Tom exists - #1321
• FILTER OUT - near duplicate - language mentioned, but not wildcard FRENCH - #3884
• FILTER OUT - near duplicate - pronoun difference only (to save time he preferred over she) - #1175
• FILTER OUT - near duplicate - PRONOUN instead of TOM - #3981
• FILTER OUT - near duplicate - simple sentence pattern - #4002
• FILTER OUT - near duplicate - various (2014-2015) - #3358
• FILTER OUT - near duplicate or likely to become one - first name not Tom or Mary - #3086
• FILTER OUT - near duplicate or likely to become one - non-standard wildcards - or Mary instead of Tom, or Mary before Tom when not necessary - #3790
• FILTER OUT - near duplicate - other - #1322
• FILTER OUT - not likely to ever use - #989 (2012)
• FILTER OUT - possibly proofread again, but don't use for now - #1262 (2013)
• FILTER OUT - Not likely to use these (2014-2015) - #3254
• FILTER OUT 4f- not likely - #1080
• FILTER OUT - English Sentences with Audio That I Don't Use - #4699

4. Sentences I have just ignored, since I'm unlikely to ever use them.

• FILTER OUT - read and postponed for some reason - #3608
• FILTER OUT - not proofread - Dima's Adventures by FeuDRenais - Many of these sentences are inappropriate for my projects - #1062
• FILTER OUT - not proofread - near duplicate or likely to become one - Amastan's non-wildcard names - #3751
• FILTER OUT - not proofread - near duplicate or likely to become one - Spamster/Hybrid's with he, she, him, her, etc - #1198
• FILTER OUT 3 - left unadopted for now - #922
• FILTER OUT 3b - left unadopted B and temporarily not use - #956 (2012)
• FILTER OUT 3c - left unadopted C - #957
• FILTER OUT 4b - ! - Ignored for now - Maybe OK, but ! is often over-used by non-native speakers - #962
• FILTER OUT 4c - ... - at least temporarily don't use these - ... is often misused - #973
• FILTER OUT for now - check again - possibly use - #914
• FILTER OUT for now - check again--corrections suggested - #913
• FILTER OUT- at least temporarily not for my projects - #649 (2011)

5. Sentences I am very unlikely to ever use.

• ! #03 FILTER OUT - Probably NEVER use these (2016) - #4680
• ! #09 FILTER OUT - perhaps NOT APPROPRIATE FOR CHILDREN -- WARNING: Some of these may not be good English sentences. I didn't spend time proofreading them. - #888
• FILTER OUT - incorrect or better exists - #934
• FILTER OUT - Probably NEVER use these (2014-2015) - #3255
• FILTER OUT - probably not use - 2013 - #1263
• FILTER OUT 6 - probably never use - #921
• FILTER OUT Strange - #1185
• FILTER OUT - Sept 2014 - not-adopted-GB88 because I don't want to use them in my projects - Some sound strange, some may be OK, but are near duplicates. Some seem to have be copied from other websites and I don't want to use copyright-infringing sentences in my projects. - #3800
• FILTER OUT - Copyright Infringement, or possibly so. Don't use. - #3892
• FILTER OUT - likely not useful things added by Hybrid from PD stuff - #4001

6. Sentences I am very, very unlikely to ever use.

• ! #10 FILTER OUT - not proofread - mention of TATOEBA or usernames and things not useful offsite - #312
• ! #11 FILTER OUT - not proofread - too long to be useful - ignored - #926
• FILTER OUT - not proofread - likely too long to be useful - #971
• FILTER OUT - not proofread - longish unadopted things ignored for now - #1176
• FILTER OUT 7c - longish - multiple sentences - #972

About

• Briefly, these are lists I make for myself. I use the "whitelisted" sentences in my own projects. I don't use the "blacklisted" sentences.
• I have sorted the "blacklisted" sentences into multiple lists, so other developers can make their own decisions on what sentences they would like to review for possible use on their own projects.
• In the past, as I proofread English sentences in the Tatoeba Corpus, I sorted them into lists.
• Recently, each week I've been proofreading the new English sentences that have been added. I read the ones by native speakers first, and if I have time, I read ones by non-native speakers, too. I have been adding the ones I want to use to the #01 WHITE LIST, and have been ignoring others since it takes more time to sort the sentences I don't want to use into lists.
• If someone can suggest a more efficient way for me to eliminate the bad sentences and harvest the good sentences, please send me a message.
• The lists I am currently most likely to be adding to are prefixed with an exclamation mark followed by a number.
  For example, ! #03 and ! #11.

Thoughts on an Effective Rating System

Use one-digit numbers 0 through 9, with 9 being the highest rating.

9 = Very likely to be good and natural-sounding.

Sentences with audio

[This level could possibly be automatically assigned.]

8 = Likely to be good and natural-sounding.

Sentences tagged OK by a native speaker

Sentences rated OK by at least one native speaker and not rated "not OK" by anyone else.

[This level could possibly be automatically assigned.]

7 = Likely to be good and natural-sounding.

Sentences by native speakers, but not rated and/or not tagged OK.

[This level could possibly be automatically assigned.]

6 = [for future use, if needed]

[for future use, if needed]

5 = Understandable and grammatically correct, but perhaps not what a native speaker would say.

If a non-native speaker used this, the intended message would very likely be understood.

[This would be a manually-chosen level by members]

4 = Unsure, but perhaps correct.

[This would be a manually-chosen level by members]

3 = Unknown. It would be safest to assume that these may not be good.

Sentences by non-native speakers that are not rated and/or not tagged OK.

[This level could possibly be automatically assigned.]

2 - Unsure, but perhaps wrong.

[This would be a manually-chosen level by members]

1 = Definitely Bad

[This would be a manually-chosen level by members]

0 = Ignored. I've read it, but I don't want to rate it now.

I don't want to look at this sentence again until I've reviewed all the others.

[This would be a manually-chosen level by members]