Name of the new format

[krischer]

We do need to decide on a name for the new format and it is also becoming a bit awkward to keep calling it NGF (Next Generation Format) in the discussions. The obvious candidate is "miniSEED 3" but some people voiced concerns against that choice.

Thus please comment if ...

1. ... you are okay with the name "miniSEED 3".
2. ... you oppose the name and why.
3. ... you have a proposal for a different name.

I'll summarize the discussion at the end of the month and then we can just vote on it.

[crotwell]

I don't really care, but if I had to pick I would say this is a brand new format, not a simple revision of miniseed, so might as well give it a new name. That makes it clear that there is no expectation of binary read/write compatibility. And having a new name also helps to ease people's fear of the existing miniseed2 going away. With a new name there is more comfort with the idea of both existing in parallel for however long is needed.

I would suggest "Novel Geophysical Format" and we can abbreviate as NGF!
(just kidding).
Or Geophysical Research Exchange, Archiving and Transmission format, aka the GREAT format! (just kidding).

[krischer]

I personally feel that "miniSEED 3" is the only reasonable choice as its the only name people will recognize. Otherwise I fear that there will be a battle for name recognition and people will not know what we are talking about. With "miniSEED 3" they will think its the new miniSEED and instantly know what to expect.

And even if the format might be different its basic characteristics like the record based structure and being "what you get from datacenters" will largely be the same.

[chad-earthscope]

I've heard some good arguments for having a new name: it will be an incompatible new format and that there is potential for confusing users when their "miniSEED" tool does not work with the "miniSEED" they got from a data center. I am sympathetic to these arguments.

On the other hand, there are good reasons to continue to call it miniSEED:

• as @krischer wrote, there is a lot of branding associated with miniSEED, folks know what to expect. With a completely different name there will also be confusion and explanation that it's the "new miniSEED".
• it is an evolution of miniSEED; a major version change communicates non-backwards compatibility.
• I expect that many miniSEED tools will be able to read both the old miniSEED and the next generation, such that users do not care or even know what the version is. This is the plan with the tools created at the IRIS DMC.

I like miniSEED 3.

[ozym]

Looking at the SEED Manual (2.4) there seem to be a number of references to both MiniSEED as well as miniSEED. The white paper about the format's future (July 2017) uses mini-SEED in its title, and there is also reference to mseed3 (e.g https://github.com/iris-edu/mseed3-evaluation) .

I'm assuming from the previous comments that there should be no initial capitalisation and the hyphen and mseed3 will be dropped for any documentation if this is accepted.

There is also reference to miniSEED2, does this imply "miniSEED3" (as opposed to something like "miniSEEDv3", or "miniSEED 3").

A rose by any other name ...

[jmsaurel]

For me, one of the particuliarities of the miniSEED is that it uses fixed size blockettes and blocks, which looks like something very related to the use of tape storage technology with sequential read.

If the NGF doesn't use any more those concepts, my opinion is not to name it miniSEED3.
If we still have fixed size bricks inside the format, then it could still be named miniSEED3.

• I expect that many miniSEED tools will be able to read both the old miniSEED and the next generation, such that users do not care or even know what the version is. This is the plan with the tools created at the IRIS DMC.

@chad-iris ,so for example, tools like msi, msrouter or msmod would continue to work with the NGF ?
In that case, we would need to keep miniSEED name, because this is implied in the tools names.

[chad-earthscope]

@chad-iris ,so for example, tools like msi, msrouter or msmod would continue to work with the NGF ?

Yes, that's the idea.

In that case, we would need to keep miniSEED name, because this is implied in the tools names.

I wouldn't advocate retaining "miniSEED" solely based on program names, but that minor detail is one of many that illustrates an important aspect of choosing a completely different name for something that is an evolution of miniSEED and will be used in the same way.

[chad-earthscope]

There is also reference to miniSEED2, does this imply "miniSEED3" (as opposed to something like "miniSEEDv3", or "miniSEED 3").

I'm agnostic to which variation beyond that the documentation should use one consistently.

[crotwell]

The more I think about it, the more I believe that NGF is just too different from miniseed2 and really needs a new name. The fixed header is completely different and the blockette structure is totally gone, what is left to justify the reuse of the name?

I just fail to see how an impartial observer looking at the two file formats would think "yea, this one and that one are really similar". Further I worry about the confusion when users only know "my program reads miniseed" and so want to download that format. But then the choices are only "miniseed2" and "miniseed3" and they have to figure out which version they need. And in the short term it will be worse as existing software will simply say "we can read miniseed" but in reality can only read miniseed2. There will be no indication in the documentation to the user as to which is correct. Given how long research software upgrade cycles are, this will be cause confusion for years if not decades after the initial release of the new format. At least a new name documents in an immediate, clear and simple way that the two formats are in not interchangeable.

That said, I do see that an indication of the linkage with existing seed and miniseed formats is desirable. What about a name that was different, but that still indicated that it was within the SEED family of formats?

Perhaps something like "nSeed" with 'n' for new or next and also coming after 'm'? It is easier to say and type then "mini-seed-three".

I am not crazy about nSeed and would be happy if someone came up with something better, but I much prefer it to reusing the name miniseed.

[chad-earthscope]

"microSEED"? ;)

[chad-earthscope]

There will be some level of confusion no matter what, I don't think there is a solution that avoids it.

Most users do not know or care about the details of a miniSEED header, it is a lot easier to educate the relatively few folks that will actually write software for it. So I'm less worried about how the internal details look, but more focused on the function and recognition of the function.

An example of a well known data format, netCDF did not drop their name when they moved from versions 3 to 4, but version 3 readers cannot read the version 4 data model. (They did change for HDF5 but that was a scope change as I understand it.)

[jmsaurel]

Perhaps something like "nSeed" with 'n' for new or next and also coming after 'm'?

I like that.
And it could stand for nanoSeed ;) which may have some link with some reduced headers and ability to have smaller packets.

[jsaul]

How about 'notSEED'? :)

[crotwell]

I like it! :)

[krischer]

The more I think about it, the more I believe that NGF is just too different from miniseed2 and really needs a new name. The fixed header is completely different and the blockette structure is totally gone, what is left to justify the reuse of the name?

It's a brand name. And how many user's really know the details of the format in the first place? For most scientists its a format data comes in and they could care less about the details. Also the information content and use cases for miniSEED 2 and the new format are pretty much identical. So many users (assuming the tools keep working) might not actually notice much of a difference.

But the main reason to go for miniSEED3 in my book would really be that its a name people know and they know what to expect from it. Reeducating the broader community might take forever and why "waste" an established "brand"?

I personally also like nSEED but its too close to mSEED and people could just think its a typo. And nanoSEED (also nice!) might imply that it contains less information than miniSEED.

I just fail to see how an impartial observer looking at the two file formats would think "yea, this one and that one are really similar". Further I worry about the confusion when users only know "my program reads miniseed" and so want to download that format. But then the choices are only "miniseed2" and "miniseed3" and they have to figure out which version they need. And in the short term it will be worse as existing software will simply say "we can read miniseed" but in reality can only read miniseed2. There will be no indication in the documentation to the user as to which is correct. Given how long research software upgrade cycles are, this will be cause confusion for years if not decades after the initial release of the new format. At least a new name documents in an immediate, clear and simple way that the two formats are in not interchangeable.

I'm not too worried about this but maybe I'm mistaken. Most tools that can work with MiniSEED 2 files are based on libmseed and @chad-iris to my understanding will update it in a way that it can also read files in the new format. Thus most tools get an almost "free" update.