Suggestion for ageprefixes

[dendwaler]

We already have a system for ageprefixes but i think it would be easyer to always use the date of release as an age prefix.
Make the last export in the format year-month-day.
Thats 6 digits and fits into the current system and does not interfere with cyan prefixes.
Its unlikely that two writers will release an age the same day therefore Agenumbers will be unique.
In the rare case that day was already a release, then simply use the next day.
The benefit of it is that you build up a library of releases in a chronical time order automaticly.

For my internal building i use as prefix the day when i start to build.
The difference of the final release and the " build under construction" shows me how long it was in delopment.

What you think?

[diafero]

I don't think this will work.

First of all: By date of release, you mean date of first release, or date of most recent release? The sequence prefix of an age should never be changed, so I assume it is first release. But how do you even know the date of first release when you start building? So it must be more like what you said later, date when building is started.

Even then, I am pretty sure we would see collisions much too frequently. It's not as unlikely as you may think - this is an instance of the famous birthday paradox. If you have 23 people in a room, the probability of two of them having their birthday at the same day is 50%. We have had way more than 23 age registrations per year, averaged over the last five years, so we could be almost entirely sure of a collision by now. Note that I count age registrations, not release - even unreleased ages need their prefix registered, because otherwise you (as author of the unreleased age) will be in trouble if some age appears in UAM with the same prefix. You won't be able to install it.

If the sequence prefix was 128bit long, we could think about using random values. This scheme is used frequently in the web, to obtain unique identifiers without central registration. However, the 24bit we have available in Uru:CC are way too short for this.

I wonder what problem it is that you are trying to solve. Do you consider age prefix registration too complicated? There's even a nice web frontend for it, and if someone has trouble they can just create a thread here, or PM me, or email me, asking for someone to register the prefix in his stead. This is all documented in the wiki. Suggestions how to make registration easier are appreciated

EDIT: Documenting the chronological order of releases, or the time of the internal development process, is on no way related to finding a unique 24bit identifier for each age. I see no reason to try to solve these two independent issue with the same approach. The chronological data can be extracted from the UAM, for example - just have a look at its RSS feed.