Closed Glitchy-Tozier closed 2 years ago
If I understand you correctly, it would lead to a somewhat stricter, but also more intuitive placing of "similar" letters. I like the idea.
If it es possible to add this functionality on top of the existing one and make the "addition" configurable, I would be happy about a corresponding pull request.
If I understand you correctly, it would lead to a somewhat stricter, but also more intuitive placing of "similar" letters. I like the idea.
That's what I was aiming for. :)
I can work on implementing this.
I can work on implementing this.
1. What is your opinion on that thought, as an addition besides pure symmetry: "_by similar positions, I mean something like [k&g in KOY](https://dl.neo-layout.org/grafik/bilder-einzeln/flat/koy-1-tkl.path.svg)_"
I would rather see k and ß in koy as symmetric (different hands, same fingers)
2. Could you give me a quick run-down on what things you think I should know to implement this as well (and as easily) as possible? I've not touched evaluation at all, so far, which is why that part of the project seems very chaotic to me.
This will probably belong to the asymmetric_keys
layout metric that you can find here:
layout_evaluation/src/metrics/layout_metrics/asymmetric_keys.rs
. The Parameters
struct is represented in the as a corresponing section in the evaluation_config.yml
. So if additional configuration is needed, add it there.
The interface of the metrics is defined as the LayoutMetric
trait in layout_evaluation/src/metrics/layout_metrics.rs
.
I would rather see k and ß in koy as symmetric (different hands, same fingers)
I didn't mean I was planning to replace the symmetry-metric. It's difinitely not ideal, but wouldn't you agree that this is a step-up from random placement?
This will probably belong to the
asymmetric_keys
layout metric that you can find here:layout_evaluation/src/metrics/layout_metrics/asymmetric_keys.rs
. TheParameters
struct is represented in the as a corresponing section in theevaluation_config.yml
. So if additional configuration is needed, add it there.The interface of the metrics is defined as the
LayoutMetric
trait inlayout_evaluation/src/metrics/layout_metrics.rs
.
Thank you, I'll start working on this sometime the next ~days~ weeks.
I would rather see k and ß in koy as symmetric (different hands, same fingers)
I didn't mean I was planning to replace the symmetry-metric. It's difinitely not ideal, but wouldn't you agree that this is a step-up from random placement?
I am not sure about it being a step-up. I would prefer an "unconstrained" placement that leaves the optimization more room to address the other metrics than a non-intuitive placement.
This will probably belong to the
asymmetric_keys
layout metric that you can find here:layout_evaluation/src/metrics/layout_metrics/asymmetric_keys.rs
. TheParameters
struct is represented in the as a corresponing section in theevaluation_config.yml
. So if additional configuration is needed, add it there. The interface of the metrics is defined as theLayoutMetric
trait inlayout_evaluation/src/metrics/layout_metrics.rs
.Thank you, I'll start working on this sometime the next days.
Ok, I'll skip on that thought for now!
I started thinking about this when I saw you removed
["gbd", "kpt"]
from the metric. On one hand I think those letters are related and thus should be grouped with each other. On the other hand I understand that it is unnecessary to haveb
→p
be in the same relative positions asg
→k
. With those non-diacritics-letters I'd actually expect them to be in "related positions". Thus I propose improving the evaluation of similar letters:1. Related position
This metric checks whether the specified letters are in sensible related positions. Something like this:
k
&g
in KOY)))2. Position-groups
This is the same as what the current metric does, according to my understanding. Symbols of specified groups shall have the same relative location to each others.
This change allows us to
aä oö uü
and also allows us to improve placements ofgk pb dt mn
without blocking too many of the other metrics.aou äöü
are grouped in clusters.