On certain browsers, there seems to be a problem where the word wraps around so that it appears as four letters instead of five and creates an unnecessary scroll bar. This is due to a problem in the margin styling for each of the pages and needs to be fixed.
Currently, during the thirty-minute waiting period from midnight EAT to 12:30 AM EAT, there is no indication that a new word has not been generated and that the old word has still not changed if this is indeed the case (e.g. GitHub Action is delayed by 5 min). In these cases, the cookies have also been reset so there is no way for the user to know whether they are playing with a new word or the same word as yesterday. Therefore, during this period (and given the above circumstances), there should simply be a notice telling users to revisit the site in 30 minutes (countdown) and informing them of the word generation schedule.
Currently the only implemented trackers are for visits and wins, which are stored on a Fauna database per the last release. However, it would be great if we could also track user interaction to see how many people have attempted to guess the word and thereby calculate statistics like success rate in the future.
As the standard 5-letter Wordle size is somewhat more difficult for Amharic, there should be one with only 4 (or even less) letters to make guessing easier and expand the list of answers/valid words. Even though 5 was chosen to allow for the most valid words, many of those words are derivatives of others and thus the actual number of 5-letter words in Amharic is likely much smaller than the current dictionary or answers dictionary size. Additionally, something similar to Contrarreloj in Spanish Wordle would be nice to have (also with four letters as five would be much too difficult with the added time pressure).
By assigning each word or day a unique ID, timezone differences can be effectively mitigated and the data pruned so that specific days or words can be selected in Fauna.
Mentioned in #20, this will utilize a query in the Fauna database to get crowdsourced statistics about user behavior and return it in the statistics panel. This is being postponed for a further release.
Current source dictionaries and their variants (dictionary.txt and answers.txt) are insufficient to validate and choose new five-letter words. While the process of validating such words may take some time to be solved (as it involves checking against verb conjugations which may not appear in the dictionary), finding an answer word list with only valid Amharic lemmas should not be too difficult. For this, we will need to investigate other Amharic word lists/dictionaries and try to find the best one.