Name brainstorming

[ecboghiu]

Hi,

I wanted to get some opinions on the best name for this package. To be short, my favourite one currently is "daginflation". But I'll write all the ones I've heard so far and my opinion of them:

• "UnifiedInflation". The original name, but I think its a bit long, and not clear unless people already know there's 2 kinds of inflation.
• "inflation". The one that sounds the best, but it is too general, and it competes with financial inflation and cosmological inflation.
• "inflaPYon". What I don't like is how it sounds when saying it aloud. Even after weeks, I still don't like it. It has the same problems as "inflation" in that it might be too general.
• "infpy". This is very short and rolls of the tongue easily. It is already taken by an EMPTY PACKAGE, such a waste of package namespace!
• "causalinflation". It is currently the name on the code on the repository. It is quite descriptive in that it brings your mind to CAUSALITY. It is a bit long though.
• "causalinf" is the short version of "causalinflation", but it competes with "inference".
• "DAGInflation". I like about this that it makes it clear we're using the DAG framework to model causality, and also it is general enough to be able to incorporate both kinds of inflation eventually.

Feel free to add your own idea or comment.

[apozas]

Given that eventually we will have both classical and quantum inflation here (and potentially even non-signaling inflations), my votes go for names that can accommodate all of them and are not too technical. These are inflapyon (sounding ɪnˈfleɪpiən, you can hear how it would sound here) because it combines the name of the technique with the information that is a Python package (and we can easily make people get used to import inflapyon as ip). Or causalinflation, because it merges the technique and the goal. For the rest, I share the cons that Cristian outlined, and in addition I don't think DAGinflation is a good name because the use of DAGs is a "technical detail", that does not give information on what the package does or is for.

[apozas]

Last time to make your voice heard about this. I want to have a last discussion about inflapyon (sounding ɪnˈˈfleɪpiən, you can hear how it would sound here). Do you like this or causalinflation? @eliewolfe @ecboghiu

[eliewolfe]

I'm not a fan of inflapyon, sorry. On the existing list, I'm partial to 1) inflation 2) causalinflation 3) causalinf What about causinflation? Or pynflation? Or pyinflation?

On Fri, Oct 28, 2022 at 12:00 PM Alex Pozas-Kerstjens < @.***> wrote:

Last time to make your voice heard about this. I want to have a last discussion about inflapyon (sounding ɪnˈˈfleɪpiən, you can hear how it would sound here http://ipa-reader.xyz/). Do you like this or causalinflation? @eliewolfe https://github.com/eliewolfe @ecboghiu https://github.com/ecboghiu

— Reply to this email directly, view it on GitHub https://github.com/ecboghiu/inflation/issues/43#issuecomment-1295171239, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACR74BEXUZNWJYOTC2IJKW3WFP2DRANCNFSM54MEFJCQ . You are receiving this because you were mentioned.Message ID: @.***>

[apozas]

I'd be happy with inflation, but there already exists a pypi package with that name. I'd thus stick with causalinflation.

[ecboghiu]

• inflapyon I like how its written, but I'm not a fan of how it sounds when saying it aloud.

• causalinf I think would be a good name, but I think "inf" can also stand for inference.

• causinflation This makes me think of "because->cuz" and not causal.

• inflation: I don't have a formed opinion yet on what I'm about to say, but I'll play a bit the devil's advocate :)

  Should we really say goodbye to a good name only because it's taken on pypi? It's elegant to run "pip install causalinflation" but this is relevant once at the beginning, whereas the package name the user will have to interact with whenever they run the code. I also think we will not lose potential users if to install people have to run "pip install git+https://github.com/ecboghiu/inflation.git@main" instead, given that we're the only package on the market, and the ease of use we provie is very attractive. People can just copy and paste the install command, or download the source code and run "pip install ." in the folder. If we also upload the package to conda, on conda-forge the "inflation" name is actually available!

  Having inflation.classical, inflation.quantum inflation.causal would be quite nice!

  I contacted the author of the pypi package but he didn't reply ... its actually an empty package. According to PyPi we can actually ask them to give us the name https://peps.python.org/pep-0541/. However, 1) its not guaranteed this will happen 2) its not super clear how quickly the resolution would occur and 3) they require info about what package we want to upload instead, and we need to send a link to the code, which we cannot do while it is private.

• between pyinflation and pynflation, I prefer pyinflation, but I would prefer even more just inflation.

• causalinflation I think this is okay. instead of "inflation.causal" I guess we would have to do "causalinflation.ns" or "causalinflation.gpt" however.

[apozas]

OK, let's then reduce the search to two options: inflation and causalinflation. About the first, some thoughts:

• If we like it, we have very many chances of obtaining the name on pypi. The current project under that name is obviously an unmaintained template submission.
• Although desirable, we do not need to have the library on pypi at release time. Thus, if we need to provide proof of existence, we can make the repository public and then upload to pypi and conda-forge simultaneously.
• I'm always concerned that people confuse our inflation with cosmological inflation. But I concede that this is not necessarily our problem.

With all this, I want to say that I am open to go through the process of claiming the name, and we have high chances of being successful. If this requires that at release time we have it only on conda (or in none), that's OK.

[ecboghiu]

When reading some of the paragraphs in the paper I have to admit it looks a bit funny to have "inflation" refer to the technique in some places, and the package name in others. E.g.:

Here we present inflation [25], an open-source library, written in Python, that imple- ments the inflation framework [20, 21] for causal compatibility. [...]

Do you think there are any problems with this, or when speaking in person about the package and referring to both the technique and the package?

[eliewolfe]

Typesetting is critical here. The package name should always be written is textsf (sans serif). That should remove any ambiguity, no?

On Thu., Nov. 3, 2022, 6:54 p.m. Cristian, @.***> wrote:

When reading some of the paragraphs in the paper I have to admit it looks a bit funny to have "inflation" refer to the technique in some places, and the package name in others. E.g.:

Here we present inflation [25], an open-source library, written in Python, that imple- ments the inflation framework [20, 21] for causal compatibility. [...]

Do you think there are any problems with this, or when speaking in person about the package and referring to both the technique and the package?

— Reply to this email directly, view it on GitHub https://github.com/ecboghiu/inflation/issues/43#issuecomment-1302767008, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACR74BB4CIXM6XSZTNBQ3K3WGQ7D5ANCNFSM54MEFJCQ . You are receiving this because you were mentioned.Message ID: @.***>

[apozas]

Yes, it should. Since this has been settled, I'm closing the issue.