Dzertanoj's diary

Recent diary entries

Non-English keywords in English iD interface.

Posted by Dzertanoj on 18 January 2018 in English (English)

I just noticed an interesting thing. If you want to create a point indicating a garbage dumpster (trash container or whatever similar) using iD with the English language interface, you add a point geometry and then start typing "garbage" in the Search field. Once you've typed "gar", found tagging options will be related to anything "garden" and "garbage", but there will also be a "ॐ Hindu Temple" as a sixth item. If you type one more letter and make it "garb", you'll have everything "garbage", but Hindu Temple will jump to the second place.

I'm not claiming that I know how it works, but I assume that there are keywords tied to an interface language and to every tag or set of tags for a specific object. They seem to be language-dependent (while it is still possible to type something in German, like "wald", and get Wood as an option), otherwise, there would be a lot more confusion with similarly spelled keywords from different languages.

However, I don't see any logic in this specific situation. As Wikipedia says, "Garbhagriha" is a Sanskrit word for a part of a Hindu temple (not even for a temple itself). How often might it be typed in Latin alphabet by iD users? Does it really belong to English interface, keeping in mind that English is not the main language in any of those countries where Hindu religion is prominent enough? It definitely belongs to Hindu and some other languages, but English? By the way, if you try typing "altar" (also a part of a temple in many religions) - nothing related to temples or cathedrals will pop up. Considering this logic, Hindu Temple should not show up on a list after typing "gar" or "garb" in a search form.

Just to be clear: I don't care if it will continue popping up - it looks amusing to me, not just "wrong". I'm not going to dig into iD's complicated structure and register on any translation services to fix that.

Communication channels

Posted by Dzertanoj on 29 December 2017 in English (English)

Being a distributed crowdsourced project, OSM strongly depends on communication channels to bring the desired level of coordination within communities. However, since there is a high level of independence and autonomy in these communities, it might lead to a situation of scattered and fragmented communication. I'm not proposing anything here, but I still want to bring this issue to attention.

Every technology used for communication (mailing lists, forums, messenger channels, IRC) might have a certain advantage over others, but the thing is that it has nothing to do with its popularity. The dominance of a certain channel within a community has historical roots and relies on a habit. An existing choice might have a justification, but that is not a reason for a particular choice. For example, why Russian-speaking community prefers forum over the mailing lists? Just because several people started using it many years ago. It's not because it's easier to search through it or because it has better message formatting features. But there is nothing wrong with it. However, there is one significant bad side.

Commonly used communication channel might get a "fork" just because several people (or even a single person) want to take over it or just because they have different habits. That's exactly what happened in Russian-speaking community starting a Telegram messenger channel. Those who don't want to use Telegram or who prefer non-real-time communication (forum or mailing list) over a real-time chat technology, become deprived of a significant part of a communication process.

Moreover, "forks" like that might be used to make certain decisions look more legitimate. For example, a certain tagging practice or a change of a Wiki page might be discussed in a limited circle of people using a new communication channel. Others, who use a different channel, stay unaware of that. While someone who participated in that discussion could later use an argument that something has been discussed and there was no objection, so let it be this way. This is a practical example of using an existing selection bias to justify something because it is easier to pick a channel of communication where a certain opinion prevails and to get an approval of any idea that correlates with that opinion (also, look up "Gerrymandering" in Wikipedia).

I am perfectly aware of the fact that there is no real approval procedure for any kind of decisions forming a structure of the OSM project. But it doesn't mean that using technicalities like that for "political" matters is something totally acceptable. I hope that someone will, at least, think about it for a minute.