Talk:Alphabetisch sortieren

At the moment, all special characters seem to be placed at very end of the alphabet, with „Sées“ after „Subiaco“ and „České Budějovice“ after „Zwettl“. As we are a multilingual project dealing with place names in many languages and serving a small but very international scholarly community, there is no obvious solution to the question of alphabetical order. Maybe that is in itself an argument for a primitive but straightforward solution: treat all accented letters, Umlaute, and the like equivalent to the „normal“ letters, so ç = č = c, é = e, ø = o, ü = u, etc., and assign sortkey accordingly. Not an urgent task, but worthwhile. atb Christof Rolker (talk) 09:01, 16 May 2025 (CEST)

I tend to agree with this. At least for now this seems like a good way to produce predictable and intuitively comprehensible results easily. While I was studying in Tübingen I worked part-time for a software company. One of my bosses was actually involved in the process of formulating the normative "European Ordering Rules". If we want to follow them in detail, it might get messy pretty soon. But, as you said, we are serving an international scholarly community, so we should at least consider this.

One thing seems clear to me: This cannot be done by a simple switch in the Wiki software as the software cannot possibly take into account the varying languages we deal with. Thus, we have to work via the sortkeys. In the (very) long term we will be able to modify the bot that assigns the sortkeys in a way that it implements the European Ordering Rules (by simply "telling" the bot, which library is in which country).

But for the time being, I'd implement your suggestion (i. e., reduce every diacritic letter to its basic latin form). Of course, this means that I will have to force overwrite all existing sortkeys in a manual run of the bot. Before I do that, I'll have to have a look at the List_of_manuscript_keys and probably fix some of the problems marked in red. After the manual run of the bot, we can start addressing individual problems, e. g., if we feel that we need to move "Č" to the end of the alphabet, we can implement it, or with German names we might want to treat "ö" like "oe" instead of like "o" and so on... Clemens Radl (talk) 11:24, 16 May 2025 (CEST)

thanks so much in advance. overwriting is not a problem - even where the bot is not perfect, it produces consistent results, and in the past was normally much better than human users (specifically, me). atb Christof Rolker (talk) 11:32, 16 May 2025 (CEST)