Normalization of Sponsors and Speakers names

[cmaureir]

At some point, we will have both sponsors and speakers with different names, so we might need to either have a tool to do a normalization, or a different approach. For example, when parsing the data, we might have 'Microsoft' and 'microsoft' or with people we might have 'Cristián Maureira-Fredes' and 'Cristián Maureira Fredes' 👯

Any ideas? maybe a script that finds similarities in names?

and also... for some european countries, we might have two people with the same name, it's a very tricky corner case, but we could have two "Michael Müller"

[anivorlis]

Thanks for opening the issue. I was thinking about the same issue.

For sponsors it should be easier, and even possible to do manually, by for example unique sorting. Speakers would be more work.
A script that find near matches could work, but manual confirmations would still be needed.

I was thinking of registering sponsors in a master file, and having references on the jsons but AFAIK json doesn't allow for circular referencess. A relational database would solve this but it would make it more difficult to contribute.

[mjmolina]

Related to this topic. I found that some had different names
An example is Visotech but it is Trayport ttps://ep2019.europython.eu/
I am considering to create another json file with the different names by company, and then we can decide how we manage/use this information.
We can follow the same structure as we use for the sponsors files.

Any thoughts? Ideas or suggestions? :)

[cmaureir]

I'd be inclined to replace old names by new ones to avoid complications, because it might tricky then to update the info.