Skip to content

at location#13

Open
solpahi wants to merge 2 commits into
balningau:masterfrom
solpahi:patch-5
Open

at location#13
solpahi wants to merge 2 commits into
balningau:masterfrom
solpahi:patch-5

Conversation

@solpahi

@solpahi solpahi commented May 28, 2014

Copy link
Copy Markdown
Contributor

Summary:
Remove "at location" places in words about buildings/structures like {canga} or {muzga}.

Detailed description:
The "at location" places in those words are unnecessary and cause the more useful places to be pushed to the right. "At location" can be expressed using {bu'u}/{sedi'o}/{tu'i} or by simply using a predicate {ti ckule gi'e diklo X} or {gi'e zvati X} or {gi'e se stuzi}.

Impact:
The negative impact of this change is relatively small. Looking at the corpus, one can see that these places are almost never used. Furthermore, in the case of {ckule} and {cange}, there are usages that incorrectly use the old x2 assuming it is the old x3.
I count the following usage numbers:
briju3 = 12,
cange2 = 10 (with at least 1 wrong usage),
ckule2 = 42 (with several using x2 as the taught subject),
malsi3 = 2,
ginka3 = 0,
jinto3 = 0,
muzga3 = 1,
tumla2 = 4,
xotli2 = 3

This shows an overall extremely small usage percent of these places. In the case of {ckule} and {cange}, it moves the more useful places further back requiring people to perform awkward place skipping. Thus the positive impact would be to facilitate the use of the more often needed places while at the same reducing the overall number of sumti places to learn.

solpahi added 2 commits May 28, 2014 15:07
Summary:
Remove "at location" places in words about buildings/structures like {canga} or {muzga}.

Detailed description:
The "at location" places in those words are unnecessary and cause the more useful places to be pushed to the right. "At location" can be expressed using {bu'u}/{di'o}/{tu'i} or by simply using a predicate {ti ckule gi'e se diklo X} or {gi'e zvati X}. 

Impact:
The negative impact of this change is relatively small. Looking at the corpus, one can see that these places are almost never used. Furthermore, in the case of {ckule} and {cange}, there are usages that incorrectly use the old x2 assuming it is the old x3.
I count the following usage numbers: 
briju3 = 12, 
cange2 = 10 (with at least 1 wrong usage), 
ckule2 = 42 (with several using x2 as the taught subject), 
malsi3 = 2,
ginka3 = 0,
muzga3 = 1,
tumla2 = 4,
xotli2 = 3

This shows an overall extremely small usage percent of these places. In the case of {ckule} and {cange}, it moves the more useful places further back requiring people to perform awkward place skipping.
@adamlopresto

Copy link
Copy Markdown

Of the words listed in the Impact section, {briju}, {malsi}, {ginka}, {muzga} and {tumla} all have the "at location" place as their last numbered place, so the argument about it pushing more important places further right doesn't stand up. {tumla} in particular is nothing but a relationship between an area of land and its location, so the location seems utterly crucial to it. {tumla} just isn't like the rest of those.

For those that do have later places, how many uses have you now invalidated? How many times has the old ckule3 been used, now incorrectly?

@solpahi

solpahi commented May 28, 2014

Copy link
Copy Markdown
Contributor Author

I agree about {tumla}, so let's remove that word from the discussion.

ckule3 has been used 77 times, 72 of those times on IRC. Excepting a few cases where ckule3 was used in the sense of ckule4, those ~70 sentences would become invalidated. However, I don't think it's a large number. Also, I don't know what you think about rewriting parts of the corpus in cases like these. The number is so low that they could be updated by hand.

@dustinlacewell

Copy link
Copy Markdown
Contributor

I agree that the low usage and the various ways to supply the same information warrants the simplifying of the predicates.

@selpahi can you please add the changes to the indvidual gismu definitions to this pull request?

@xorxes

xorxes commented May 28, 2014

Copy link
Copy Markdown

Supporting fact for removing "at location" place from those: gusta, barja,
zarci, banxa, dinju, zdani don't have one.

I'd be very happy if location and time places were removed from "jbena",
too. This may be the only predicate with time and location places.

@solpahi

solpahi commented May 28, 2014

Copy link
Copy Markdown
Contributor Author

la .xorxes. cu cusku di'e

Supporting fact for removing "at location" place from those: gusta, barja,
zarci, banxa, dinju, zdani don't have one.

Good point.

I'd be very happy if location and time places were removed from "jbena",
too. This may be the only predicate with time and location places.

ie sai sai

@durka

durka commented May 28, 2014

Copy link
Copy Markdown
Contributor

Isn't there a relatively high rate of usage of {terjbe} for "birthday"?

On Wednesday, May 28, 2014 at 5:21 PM, selpahi wrote:

la .xorxes. cu cusku di'e

Supporting fact for removing "at location" place from those: gusta, barja,
zarci, banxa, dinju, zdani don't have one.

Good point.

I'd be very happy if location and time places were removed from "jbena",
too. This may be the only predicate with time and location places.

ie sai sai


Reply to this email directly or view it on GitHub (#13 (comment)).

@solpahi

solpahi commented May 28, 2014

Copy link
Copy Markdown
Contributor Author

la durka cu cusku di'e

Isn't there a relatively high rate of usage of {terjbe} for "birthday"?

Just 33 uses in the corpus.

Also, I think {jbedetri} (so far 15 uses) is better, since you only have
one day on which you are born, but celebrate on every day with the same
date label.

@xorxes

xorxes commented May 28, 2014

Copy link
Copy Markdown

Having a time place makes using "jbena" with tenses very very awkward.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants