This post is from a low-reputation account and contains an unverified outbound link. Be cautious before clicking external links.

@grammarnazi: Recent Updates

grammarnazi(5)

Published in

#programming

Words

651

Reading

3 min

Listen to this post

~4 min

Grammarnazi Header 2.jpg
A week has passed since the introduction post, which resulted in a lot of constructive feedback. @omitaylor provided a whole list of ideas, and a post by @kus-knee gathered almost 100 comments worth of intel. Below is a summary of changes which have been initialized during this last week, as well as a few plans for the future.

Upgrades made since the previous post:

The typo dictionary was extended:

Stremit will be corrected to Steemit.
Chasig will be corrected to chasing.
Otder will be corrected to order.
Fith will in certain cases be corrected to with.
Sencond and senconds will be corrected to second and seconds, respectively. Fun fact: This typo was actually present in the previous post.
Whic will be corrected to which.

The grammatical mistake dictionary was extended:

In the long term will be corrected to in the long run.
Good will in certain cases where it's used as an adverb be corrected to well.
Without further due and without further adieu will be corrected to without further ado.

The misspelling dictionary was extended:

Filosophy will be corrected to philosophy.

Algorithmic changes:

The bot was previously set to scan through the most recent Steemit post every five seconds. When the bot was initially created, some unscientific testing revealed that the average time between each new post was about five seconds. This obviously means that sometimes posts are published more frequently, so now the algorithm has been adjusted to scan the most recent Steemit post every three seconds. The time between each scan could of course be set even lower, but some brief testing revealed that with the current setting, the algorithm almost never misses a post.

The bot has been updated to contain a list of tags whose posts it will not scan. Currently, the only posts it will not scan, are posts carrying the freewrite tag. Wikipedia defines freewriting as a prewriting technique in which a person writes continuously for a set period of time without regard to spelling, grammar, or topic.

Multiple comments suggested that the bot should stick to tags which are mainly used by fresh Steemians, to minimize the chance of getting a heavy downvote, thus not putting @grammarnazi's reputation at risk. This would, however, also remove some heavy upvotes, so for now, the bot will remain unchanged in regards to this.

The bot already avoided posting comments on posts by users who had previously downvoted it. Still, those same users might stumbe upon another user's post whose mistakes were corrected by @grammarnazi. To limit the amount of cases where this might happen, the algorithm has been adjusted to not post comments on posts which refer to users who have previously downvoted a comment made by @grammarnazi.

Other changes

Some users seem to consider the name of the bot offensive, and have presented a lot of alternatives. While it's understandable that the nazi-part of the name might be offensive to some, grammarnazi is a well-established internet term which instantly makes everyone understand its purpose, so the name will remain the same. To make the overall feel of the bot more friendly, the header image of the account now features pixelated hearts.

Ideas for the future:

Post more diverse correction comments, to make the bot seem more human. The comments could also be made friendlier.
Maybe post a statistics post every week, containing info about most frequently corrected mistakes, number of comments, upvotes and downvotes...

Delegation links:

Some comments requested links to delegate Steem Power. To delegate Steem Power, simply click one of the links below and sign the delegation transaction with SteemConnect.