Open morinted opened 8 years ago
I've removed a few thousand misstrokes I think from the main dictionary for myself where the commit history mostly shows why I felt the stroke was off: https://github.com/dimonster/plover-dictionaries/commits/master/dictionaries/dict.json
I hit some issues with conflicts and inconsistencies around these briefs:
"KPESZ": "excess",
"KPES": "access",
"KPES/S*EUF": "excessive",
"KPES/-BL": "accessible",
"SEBLT": "accessibility",
"KPES/REUS": "accessories",
"SESZ/PAOL": "cesspool",
It looks to me like:
So I'd suggest:
"SESZ"
were "access".The main briefs might then be:
"SESZ": "access",
"SESZ/REU": "accessory",
"SEBL": "accessible",
"SEBLT": "accessibility",
"KPES": "excess",
"KPES/S*EUF": "excessive",
"SES/PAOL": "cesspool",
Altogether:
"SESZ": "access",
"SESZ/-BL": "accessible",
"SESZ/-BLT": "accessibility",
"SEBL": "accessible",
"SEBLT": "accessibility",
"EUPB/SESZ/-BL": "inaccessible",
"EUPB/SESZ/-BLT": "inaccessibility",
"SESZ/REU": "accessory",
"SESZ/REUS": "accessories",
"SESZ/RAOEUZ": "accessorize",
"KPES": "excess",
"KPES/S*EUF": "excessive",
"SES/PAOL": "cesspool",
Removals:
-"SESZ/PAOL": "cesspool",
-"KPES": "access",
-"KPES/-BL": "accessible",
-"KPES/REUS": "accessories",
-"KPES/TO": "access to",
"AOELD/SRAOEUS": "Edelweiss"
"AOED/*L/SRAOEUS": "Edelweiss"
"PRO/SPEBGS": "prospection"
I'm sure that the Plover dictionary will be happy to greet that blossom of snow.
...
"TK*PB/TK*PB": "deoxyribonucleic acid"
remove "TKRESZ" for "dress"
"A/KUFPLD": "accustomed"
exists but not "A/KUFPL": "accustom"
.
"KRAO*UT": "crwth"
Here are some briefs for Canadian provinces similar to the star briefs that already exist for US states:
"KPWR*": "British Columbia" (mnemonic here is "BC" on the left side)
"A*B": "Alberta" (may conflict with lots of entries that otherwise begin with A*B)
"S*BG": "Saskatchewan" (this one conflicts with South Carolina; one of them would have to be assigned the double stroke)
"SK*": "Saskatchewan" (for "SK" more directly; I missed this one the first time around)
"PH*B": "Manitoba" (for "MB")
"O*PBT": "Ontario" (for "Ont."; "O*N" is mapped to the "-on" suffix, but "O*NT" is mapped to "on the" which is already "O-NT" without the star)
"A*UPB": "Ontario" (for "ON" but using the "AU" as a substitute)
"KW*BG": "Quebec" (for "QC")
"TPH*B": "New Brunswick" (for "NB")
"TPH*S": "Nova Scotia" (for "NS")
"P*E"; "Prince Edward Island" (for "PE"; "P*EU" is taken by "pi")
"PAO*E": "P.E.I." (because enough people pronounce it that way)
"TPH*F" "Newfoundland" (for "NF")
"TPH-FLD": "Newfoundland" (for "Nfld.")
"TPH*FL": "Newfoundland and Labrador" (for "NF"+"L")
"TPH*L": "Newfoundland and Labrador" (for "NL")
The current dictionary is very US-centric in its geographic lexicon; I hope to at least include your neighbo(u)r to the north where I live in that 😛
malicious stroke: FLIKT
→ conflict, but FLIKT/-D
→ complicated in the current dictionary
add the FR-as-m trick to some words:
"KAFRL": "camel"
"HAFRP/SHAOEUR": "Hampshire"
"NU/HAFRP/SHAOEUR": "New Hampshire"
"HAFRP/*ER": "hamper"
"OE/HREUFRP/KWRAD": "Olympiad"
"PHROFR/SR*ERGS": "weekly-v4.0.0.dev8+8.ge062cda6"
"PHROFR/STKR*ERGS": "4.0.0.20180620"
And then change it every time you update either Plover itself or the dictionary.
"PHEL/TPHROUS": "mellifluous"
"PHEL/TPHRAOUS": "mellifluous"
Add:
"HREUR/SEUFT": "lyricist"
Remove:
"SKP/HREUR/SEUFT": "lyricist"
Please add:
"TKPWRO*EUPB": "groyne"
It is a type of rock wall that prevents erosion.
For consistency sake, REF
and R*EF
should be swapped, and reflog should be switched to be written with REF
instead of R*EF
:
"R*EF": "rev",
"REF": "ref",
"REF/HROG": "reflog",
Because if you look at the distribution of default entries in the dictionary that start with those strokes, REF
has a pretty even split between ref- and rev- words, but besides reflog literally every R*EF
word starts with rev-
Some more work could definitely be definitely be done with regard to the consistency of these two strokes in the future - some R*EF
words specifically seem to be for disambiguation (as in the case of R*EF/RAOE
→ reverie versus REF/RAOE
→ referee) or just for redundancy (like with how REF/HRAEUGS
and R*EF/HRAEUGS
both map to revelation) but there are also some words that can only be written with R*EF
even though they create no conflicts (such as R*EF/KABL
→ revocable) so this is definitely something that should be revisited later (and I probably will unless someone else does it first)
Think I've come across an orthography rule issue, less a dictionary issue.
mimic: PHEUPL/EUBG PHEUPL/PHEUBG
But when you try to add a third stroke for the -ED ending you get:
mimicced <<< this is wrong, should be mimicked
Other endings seem to work properly: mimicking mimics mimicker
Here's a list of other verbs to check for the same thing: http://wordtoolbox.com/verbs-ending-with/c
Specifically, the 2-syllable words ending in an unstressed "ic": mimicked, panicked, frolicked, picnicked, bivouacked, shellacked, trafficked
The 1-syllable words in that list should not have the "k" applied to them: arced, sicced, synced
"result": RAOE/SULT
"result": RAOE/S*ULT
To be more consistent in using RAOE
as an alternative "re-" prefix for words with multiple-stroke definitions.
“ABGS”: action
“AEBGS”: axe
“A*BGS”: axe
Right now AEBGS is “acute coronary syndromes”, which is definitely a work brief that shouldn’t exist in the main dictionary.
Also, “action” strikes me as the more common word by a long shot, so I think it should get the standard stroke.
"-T/PHET": "the Met"
Should be removed or changed because it currently can cause word boundary errors, e.g. with the phrase "the metronome" coming out as "the Met ro gnome"
"-T/PHET/PWOLG": "the metabolic",
"-T/PHET/REUBG": "the metric",
Removing that would also mean that these two lines can be removed because they seem to only exist to compensate for that exact problem
I agree. I've actually been meaning to zap that entry for a while. It causes no end of trouble.
On Sat, Jul 14, 2018 at 7:12 PM, amikumanto notifications@github.com wrote:
"-T/PHET/PWOLG": "the metabolic", "-T/PHET/REUBG": "the metric",
Removing that would also mean that these two lines can be removed because they seem to only exist to compensate for that exact problem
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openstenoproject/plover/issues/400#issuecomment-405055646, or mute the thread https://github.com/notifications/unsubscribe-auth/AAMvPmDN1Xtfxd28LFLnhZsp-yNGe3yMks5uGnrAgaJpZM4HqSne .
-- Mirabai Knight, CRC, RDR StenoKnight CART Services 917 576 4989 mkk@stenoknight.com http://stenoknight.com
As discussed on Discord, I'd suggest
"TKE/PHOET/SRAEUT": "demotivate"
because right now that translates to "demote SRAEUT" instead.
Rifle/rifles and rival/rivals are crossmatched; one has the asterisk in singular and another in plural. We should swap them to make them consistent.
For "figurative", TPEUG/TEUF
should be added.
Remove SKO SO* THE
for "so they" because I can't see how that stroke would be useful.
And maybe change KWR-S
to "why is" or at least remove the space before the comma. Also, I think a better way to write "s, " would be -RBGSZ
.
Possible additions: "A/STPALT": "asphalt", "SHAOR": "ashore", "TUPL": " tuple"
You probably also want to change either TUPLZ
or TUPLS
to the plural form "tuples". And then TUPL/-S
needs to go too.
Phonetically "tuple" should be TAOUPL because it's a long U. TUPL would be "tum" or "tupple" (if the latter were a word).
Oh, I think I see the confusion now. You pronounce "tuple" like it's "tupple" (with a short U), which I've also heard people do.
"WHAOF": "whoof"
"WHAOPS/KWREU/DAEUS/KWREU": "whoopsy-daisy"
"KROD/SKW*EUPBG/*L": "crodsquinkle"
"TPROB/SKOT/*L": "frobscottle"
"TPROB/SKOTL": "frobscottle"
"TPHRUPL/*ER": "flummer"
okay but now seriously
the following consists of (hopefully) unopinioned vocabulary additions and support for the -TD-as-ded pattern to tuck -ed endings on some words.
"A/HRAOEPBLGS": "allegiance"
"A/POEUPBTD": "appointed"
"A/RAEUD": "arrayed"
"SROEUTD": "voided"
(also for avoided)
"A/SROEUTD": "avoided"
"A/TAG": "attacking"
remove explicit entry for testing and tested (now we can fold -G on attesting and attested)
"A/TEPBTD": "attended"
"A/TOEPLT": "atonement"
"A/TPEURPLGS": "affirmation"
(not my usual stroke, but makes sense to have in)
"A/TPHAOEULT": "annihilate"
remove explicit entry for visits (now we can fold -S on activists)
"ABL/TROS": "albatross"
"AEUTD": "aided"
"A/SRAPBT/TKPWARD": "avant-garde"
"AFPBT/TKPWARD": "avant-garde"
"A*FPBT/TKPWARD": "avant-garde"
"AFRB/KWRAPS": "ambiance"
"APL/KWRAPS": "ambiance"
"AFRP/TPAOEU": "amplify"
"AFRPL/TPAOEU": "amplify"
"AFRPB/AEUPBG/*L": "archangel"
"AG/SRAEUTD": "aggravated"
(or, use SRAEUT for serrate and remove the explicit entry for serrated. I prefer adding this entry for aggravated)
"AL/KPAPBD/KWRA": "Alexandria"
"ALT/TEUF": "alternative"
(perhaps someone will also find my pseudo-brief useful)
"AO*EULD": "idle"
"AOEULDZ": "islands"
"AOURBG/KWRA": "eureka"
(a suggestion. this is my way of confronting this strange word)
"APB/HRO*G": "analogue"
"APB/TKAOEUDZ": "anodized"
"APBD/KWREU/TKAOU/TPRAEUPB": "Andy Dufresne"
"APBLGT": "agility"
(for those who "lean right", as in ARPBD for around and similar)
"ARPT": "apart"
"ART/KHRAEUTD": "articulated"
"ART/PWAOUTS": "attributes"
(or remove explicit entry for beauties)
"AUFR/KROUTD": "overcrowded"
"EBGS/KHRAOUTD": "excluded"
"EBGS/SEPB/TREUFT": "eccentricity"
"EBGS/SEPB/TR*EUS": "eccentrist"
(for people who, in conversation, occasionally can't think of the noun "eccentric")
"EBGS/EUG/WOUS": "exiguous"
"EBGT/-BL": "equitable"
(for those who don't know steno-school KWET for equity)
"EFR/SREFPBS": "effervescence"
"EPB/TEPBGS": "intention"
(for people who forget how to spell when the going gets hard)
"EPL/PWROEUD/*ERG": "embroidering"
(because embroid is not a root)
"EPLS/REU": "emissary"
"ROEUR/KWROUS": "erroneously"
"ER/ROEPBS/HREU": "erroneously"
"ER/TKPWOPL/TER": "ergometer"
"EUFRP/SRAOEUS": "improvise"
"EUFRP/SRAOEUZ": "improvise"
"EUPB/EBGS/PHREUFT": "inexplicit"
(or remove misstroke EUPB/EBGS for index)
"EUPB/KAOPBLT": "inconsistently"
"EUPB/ST*EUPBGT": "instinct"
"STPHAPBLT": "instantly"
"SPWEPBTD": "intended"
"EUT/TEUF": "iterative"
"EURT/TEUF": "iterative"
"HAEPBS": "happiness"
"HAEUPBS": "heinous"
"HAPL/PWRURG": "hamburger"
"HAS/KWREPBD": "hacienda"
"HAS/KWREPB/TKA": "hacienda"
"HEBGS/TKPWOPBL": "hexagonal"
"HR*EUL/PUT": "Lilliput"
"HRA/TEBGS": "LaTeX"
(for people who pronounce it correctly)
"HRAOER/K-L": "lyrical"
(common enough for one to slop a little and forget to use the root word)
remove explicit entry for lights (now we can fold -S on eletrolytes)
remove explicit entry for rating (now we can fold -G on liberating)
"HUB/ARD": "hubbard"
"HUB/BARD": "hubbard"
"K-BLT": "accountability"
"KABG/TPOEUPB": "cacophony"
"KAUPB/TAEUPBLG/KWROPB": "contagion"
"KAUPB/WHRAOUGS": "convolution"
"KAUPB/WHRAOUT": "convolute"
"KAURPBT/SRER/SEU": "controversy"
(for those without the brief but with the contra prefix)
"KAURPBT/SRERS/KWREU": "controversy"
(for those without the brief but with the contra prefix)
"KEP/HRER": "Kepler"
"KEPB": "ken"
(keep asterisked version for the proper name Ken)
"KHAOER/HRAOEGD": "cheerleading"
"KHAOES/PWRURG": "cheeseburger"
"KHAOEZ/PWRURG": "cheeseburger"
"KHAOUZ": "chews"
"KHRAOEUTD": "collided"
"KHRAT/*ERG": "clattering"
"KHREUFP": "cliche"
(orthographic)
"KHROR/PHRAFT": "chloroplast"
"KOEUPL/PWRA": "Coimbra"
"KPAPLGS": "examination"
(alternate brief)
"KWAOEPB/HRAOEPB": "Queen Leene"
(a pleasant surprise designed for TypeRacer aficionados)
"O*EUPBD/KREUPBL": "endocrinal"
"O*EUT/PHEU": "autonomy"
"O*URZ": "ourselves"
(common enough to concede a misstroke on the -SZ; or, for certain lazy people)
"OEB/KWREU/WAPB": "Obi-Wan"
"OFT/HRER": "ostler"
"P-RBT": "punishment"
"PAELG": "pealing"
"PAOEPB/KHRAD": "piña colada"
"PEPBT/TOPB/EUBG": "pentatonic"
"PEPBT/TO*PBG": "pentatonic"
"PEPBT/TOPBG": "pentatonic"
"PER/SKWRA": "Pera"
"PHAERBGSZ": "matrices"
"PHAOER/K-L": "miracle"
"PHAOEUPBTD": "minded"
"PHARPBLG/SA*EUGS": "marginalization"
(#savethestrokes)
"PHERPLZ": "memorize"
"PHER/KAPBT/HREUFT": "mercantilist"
"PHER/KAPBT/HREUFPL": "mercantilism"
"PHO*FR/KWRUS": "Morpheus"
"PHOPBT/TPHOUS": "monotonous"
(#strokelivesmatter)
"PHOR/ROE": "morrow"
(save asterisked version for Morrow)
"PHO*ET": "mote"
"PHROPB/EUBG": "moronic"
"POEUL/HAOERL": "polyhedral"
"PREP/TAEUGS": "precipitation"
"PREUPB/SEPS": "princeps"
"PRO/POEFL": "proposal"
(reasonable alternative to the brief)
"PRO/TOS": "Protoss"
(it already translates, might as well capitalize)
"PUBG/KA": "pukka"
"RARTD": "regarded"
"S*EUG/WOUS": "exiguous"
"SHAU/SHA*PBG": "Shawshank"
"SKETD": "scheduled"
"SORPL/SAULT": "somersault"
"SOUPBTD": "sounded"
"SPAOED/R*UPBG": "speedrunning"
"SPAOED/R*UPB": "speedrun"
"SPOPBTD": "responded"
"SR*ELTD": "veldt"
"SR*EUPL/KWRUPL": "Vimium"
"SR-RB": "vanish"
"TPAU/RAEU": "foray"
KWAOEUPB
is bound to the {^quine}
suffix, but quine is also a word. suggestions:
"KWAO*EUPB": "quine"
"KWAO*EUPB/KWAO*EUPB": "KWAO*EUPB/KWAO*EUPB"
finally, remove"STOEUPB/TAOEUP"
, avoiding quine-like typos like "stenotypos are great"
I'm a bit confused by some of the entries:
"P-P": "{^.^}", "P-P/KROPL": "{^.^}com", "P-P/O*EU": "{^.io}", "P-P/O*RG": "{^.^ org}", "P-P/ORG": "{^.^}org", "P-P/PHEUL": "{^.mil}", "P-P/SKWR-S": "{^.^}js", "P-P/TPH*ET": ".NET",
There's also:
"KROPL": "{^.com}",
So for the name of a website or for the name of a file, one wants to append a period and then the file extension or the top-level domain.
But these don't seem to be working as I would expect them to:
"P-P/O*EU": "{^.io}", "P-P/PHEUL": "{^.mil}", "P-P/KROPL": "{^.^}com",
If one is trying to do something like "person.mil" it comes out as "person.mill". Or if one wants to do "person.io" it comes out as "person.oy". Likewise although KROPL works if one wants to do "google.com" if one uses P-P/KROPL it comes out as "google..com".
These don't seem to work as expected either.
"P-P/TPH*ET": ".NET", "P-P/O*RG": "{^.^ org}",
In the case of P-P/TPH*ET it leads to "land..net" or in the case of P-P/O*RG it comes out as: "place.^ org".
So the Dictionary Format reference speaks about the distinction between "{^}ish" and "{^ish}". In the case of "{^ish}" it applies English orthography rules, while in the case of "{^}ish" it just appends the suffix.
Seems like one generally wants it not to apply orthographic rules in the case of putting a suffix onto a file or the top level domain for a website.
SEUGT
was updated to be sitting
instead of signature
in the main dictionary, but SEUGTS
and SEUGT/-S
still have defined entries for signatures
which should likely be removed as sittings
is a valid word to use.
Somme suggestions: "PRO/TPERBL": "professional", (some might try to write it like this) "PAEUFT/-L": "pastel", (phonetic) "KWRURPBD": "I understand", (analogy with briefs like KWR-PB, KWR-PL, KWROEURPBD and several others) "STWARTS": "starts with", (start with is already STWART) "TPHEPBL": "in general", (new brief)
And remove: "KARPB/SRAL/-F": "carnival", (conflict with carnival of) "SKEURB/WARB/ER": "dishwasher", (unlikely misstroke)
Edit: explained additions
"TKAOEURP": "diaper"
PR #1003 sort of made me want the following briefs:
"SPHAOBG": "mistook",
(for consistency with SPHAEUBG for "mistake")
"TPAUFL": "fall of",
"TPAUFLT": "fall of the",
"SKPROEUD": "android",
(with SKP for and)
"SPWROEUD": "Android",
(I get entroid isn't the right pronunciation, but easy enough to remember)
"RAOEUFT": "right of",
"RAOEUFTS": "rights of",
Maybe I'm folding in "of" a bit too often, speaking of which:
"SPAOEF": "speak of",
(this used to be "percieve", so)
"SPAO*EF": "percieve",
(fix that with an asterisk)
"SPAOEFG": "speaking of",
(this used to be "percieving", so)
"SPAO*EFG": "percieving",
(fix that with an asterisk)
"KW*ELT": "commonwealth"
"240": "240"
seems a bit unnecessary.
Also, not sure what's going on with "SOFT/HRUF": "SOFT/HRUF"
...
Remove "TOED/AS": "today's"
Since Plover's translation engine currently looks at the longest possible series of strokes in any dictionary to determine how far it has to look back to merge multi-stroke translations, really long stroke sequences drag down the performance of the entire program. In particular, I found these 10 entries which, if removed, reduce the longest stroke sequence in the main dictionary from 11 to 8:
"-T/PHAPB/WHO/PHEUZ/TAOBG/HEUS/WAOEUF/TPOR/AEU/HAT": "The Man Who Mistook His Wife For A Hat",
"A*/KR*/*E/T*/O*/P*/H*/*EU/HR*/*EU/KR*": "acidophilic",
"A/SAOET/*L/ALD/HAOEU/TKE/HAOEURD/SKWREPB/AEUZ": "acetaldehyde dehydrogenase",
"EUPBT/KR*/O*/TPH*/TK*/*EU/HR*/A*/R*": "intercondylar",
"HEUFP/HAOEUBG/*ER/AES/TKPWAOEUD/TO/-T/TKPWAL/ABGS/KWREU": "Hitchhiker's Guide to the Galaxy",
"HRAOUS/KWREU/TPH/-T/SKAOEU/W/TKAOEU/PHOPBD/-S": "Lucy In the Sky with Diamonds",
"KAUPB/SREPBGS/-F/RAOEUT/-S/-F/PERPBS/W/TKABLT/-S": "Convention of Rights of Persons with Disabilities",
"KR*/A*/HR*/HR*/*EU/TPH*/O*/TKPW*/*E/TPH*": "kininogen",
"TKE/KHRAOEUPB/APBD/TPAUL/-FT/RO*EPL/A*PB/EPL/PAOEUR": "Decline and Fall of the Roman Empire",
"TKO/APB/TKROEUD/-S/TKRAOEPL/-F/HREBG/TREUBG/SHAOEP": "Do Androids Dream of Electric Sheep",
These are all obscure/overly specific entries that are overwhelmingly likely to end up being removed anyway, but doing so sooner rather than later would speed up the software a little bit.
"PROS/PEBGT/HREF/TPERTS": "Prospect Lefferts"
"PHET/TPOER": "no",
strikes me as... odd. Perhaps change this to "metaphor"?"TPHOEU": "annoy",
by analogy with "TPHOEUG": "annoying",
(currently "noy," which isn't a word AFAIK) "TK-PBGS": "definition",
by analogy with "TK-PB": "define",
"TPHEBG/RO/PHAPB/SER": "necromancer"
to be consistent with the {necro^}
definition, TPHEBG/RO
And NEK/RO/MANS/*ER if it's not already there.
On Thu, Sep 20, 2018, 6:01 PM Jeremy Geels notifications@github.com wrote:
"TPHEBG/RO/PHAPB/SER": "necromancer" to be consistent with the {necro^} definition, TPHEBG/RO
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openstenoproject/plover/issues/400#issuecomment-423348432, or mute the thread https://github.com/notifications/unsubscribe-auth/AFkgSs_wYJJCb_ek-d8t6D36-eAXbXkJks5udBBRgaJpZM4HqSne .
For removal:
"*EBGS/HEUBT": "Exhibit#G",
"KPEUBT/TPHUPL/PWER": "Exhibit No.#G",
"P*BGS": "People's Exhibit No.#G",
"P-BGS": "People's Exhibit#G",
"PHR-BGS/-S": "Plaintiff's Exhibits#G",
"R*PBGS": "Respondent's Exhibit No.#G",
"TK*BGS": "Deposition Exhibit No.#G",
"TK*FBGS": "Defendant's Exhibit No.#G",
"TK-FBGS": "Defendant's Exhibit#G",
"TK-FT/AOES/TPHUPL/PWER": "Defendant's No.#G",
"TKPW*BGS": "Government's Exhibit No.#G",
"TKPW-BGS": "Government's Exhibit#G",
These entries all seem way too overly specific
Suggest to remove:
"THA/AOUGS": "that use"
As I don't see the point (is this a misstroke?) and this conflicts in a sentence like we are trying to do that using the following method.
"OERDZ": "other side"
It's currently mapped to O*ERDZ, and there's no reason it shouldn't be mapped to the unstarred version too.
Remove: "PRAOEUT/-S": "priorities"
It's a misstroke (it should be PRAOEURT/-S
but really just remove it) that clashes with PRAOEUT
for parasite
, turning the expected output of parasites
into priorities
Suggest to add:
"KAUPB/TPHEBGT": "connect"
We already have KOPB/TPHEBGT
and the brief KEBGT
for connect
, but you should be able to use the prefix stroke KAUPB
as well, I think.
Edit: Similarly:
"KAUPB/SERPB": "concern"
Remove:
"W/STAPB": "withstand",
"W/STAPB/HREU": "with Stanley",
"STAPB/HRAOE": "Stanley",
The first creates word boundary errors for the phrase "with Stan", the second is just there to fix that error for the name Stanley (I guess for some reason Stanley is a more important name than Stan?), and the third creates a word boundary error with Stan Lee, even though the more standard way to right Stanley STAPB/LEU
is also in the default dictionary
Report suggestions for changing the default dictionary in this thread
Just reply or edit posts to give suggestions, we'll remove them as they get fixed.