Hacker News new | ask | show | jobs
by thaumasiotes 2636 days ago
In this particular case, I'm guessing -- based on the URLs -- that in order to have the Instagram username "sussexroyal", it's also necessary to be located at www.instagram.com/sussexroyal . This is a design mistake. The URL identifier does not need to be the public-facing identifier, and it shouldn't be. You can have the Duke of Sussex as sussexroyal at www.instagram.com/y6llflk9 and the guy in Sussex who roots for the Royals as sussexroyal at www.instagram.com/y2q9g6uo .

There is no need to have the public-facing identifier be the unique identifier, nor is there a good reason to do it. ICQ got this right, way back at the beginning. But somehow everyone forgot.

3 comments

> There is no need to have the public-facing identifier be the unique identifier

Yeah, there is: it's a major UX improvement, because the URLs at issue are entry-point URLs, which need friendly names for the same reason domains for public-facing services (which are key components of entry-point URLs) do.

It's true that in the general case objects don't necessarily need a URL component that matches their friendly name, but this is not the general case.

It seems like a major UX improvement, but is somebody looking for Prince Harry's Instagram account going to know http://www.instragram.com/sussexroyal, or are they going to enter the app and search "Prince Harry", click the first link, and then "Favorite" that account for future reference?

I know I personally very rarely type a direct URL. Either it's already bookmarked, or I google it.

> It seems like a major UX improvement, but is somebody looking for Prince Harry's Instagram account going to know http://www.instragram.com/sussexroyal

Instagram accounts (and similar social media IDs) are often communicated in print, and parsing and accurately transcribing things that work like natural language is a lot easier than something like an arbitrary base36-encoded identifier.

They are also sometimes communicated via orally or via radio, where being able to hear and remember is even more affected by the using natural language.

But instagram.com/sussexroyal is a much more memorable URL than instagram.com/y2q9g6uo, and being able to find someone from remembering their handle is much better than remembering the handle and then wondering which of the multiple people called sussexroyal you're after when the search results come through

Needless to say, the most prominent and memorable identifier being non-unique has all sorts of uses for trolls and spambots too.

> Needless to say, the most prominent and memorable identifier being non-unique has all sorts of uses for trolls and spambots too.

Sure, but it has more uses for people who would like to have a reasonable name. This is just the "knives can be used to kill people" argument.

The option for a reasonable name still exists (and site owners don't have to arbitrarily confiscate it to give it to more PR-worthy people), it just might be a slightly longer reasonable name with a disambiguator built into it instead of a common first name or cool dictionary word, and you get a reasonable URL as a bonus. I'd probably rather be found at instagram.com/johnsmithspringfield than instagram.com/jh9fjhfgjhg (or 'search for "John Smith" and scroll through 300 entries') anyway.

Of course with unique identifiers you probably don't get to call yourself elonmusk, POTUS or amazon

> instagram.com/sussexroyal is a much more memorable URL than instagram.com/y2q9g6uo

Since there are a limited number of memorable names available, anyone late to the party gets stuck with non-memorable names. Why not even the playing field?

Most social networks would see rewarding early adopters as a benefit rather than a drawback. Besides which, TheThreeWordName or JohnBSmith1982 is still a lot more memorable than a random alphanumeric string.

As another person said further down, it's why we have domain names rather than IP addresses. Sure, the way in which they're distributed might be suboptimal, but not nearly as suboptimal as making everyone have to remember the IP address or rely entirely upon a search function that returns ever-changing results.

> Most social networks would see rewarding early adopters as a benefit rather than a drawback.

But in this particular case, they weren't the early adopters...

The real "early adopter" was robbed of his name, for a (imho) is a lame excuse.

You really think that? Why do we even have domain names then? Why don't we tell everyone to remember IP addresses?