[seqfan] Re: Can we create a workable taxonomy for classifying all the various kinds of sequences in the OEIS?

[Wayne VanWeerthuizen]

Marc, I totally agree, especially at this stage. 

To clarify again, my interest was mainly just in the initial development of a detailed comprehensive category list itself, and not in any specific modification of the OEIS to make use of the list in any specific way, whether for searching or for any other purpose. Incorporating such a list, once completed, into the OEIS itself or into some external resource, might be a discussion for a later time. I think getting into it now is "putting the cart before the horse", as they say. For now, I am still working to determine how really feasible it is, in the first place, to classify sequences into a few different fundamental kinds.

For now, I rather just wanted to discuss the various possible categories themselves. I am still in the process of revising and expanding my own list, and seeing how well it works to characterize randomly selected sequences from the OEIS. So far I have just thirteen major categories and about sixty subcategories. This is still a work in progress. I expect to add many more subcategories. But I expect only a few additional major categories will be needed, but I see what more is needed as I continue to examine other OEIS sequences.

So far, the most complicated sequence generation meta-rules I've discovered in the OEIS have involved "racing sequences", and also what I've previously called "least-k sequences". What other examples of complicated meta-rules can anyone point me to? Also, what does "convolution" mean, specifically in the context of integer sequences? (I see "self-convolution" mentioned in A000106.)


On 4/9/2018 10:37 PM, Marc LeBrun wrote:
> For this work there's absolutely no need to extend the OEIS itself, because a reference like https://oeis.org/A123456 <https://oeis.org/A123456> is a perfectly good URI.
> You can erect a classic "semantic web" using these references (say with OWL et al) as a completely external RESTful resource (or whatever).
> Moreover, a multitude of such metadata services could happily coexist, even cross-link, without requiring any impact on the OEIS.
> Building such commentaries upon, rather than within, the OEIS could be a bounteous source of interesting collaborative projects.
> Let a thousand ontologies bloom!
> 
>> On Apr 9, 2018, at 11:06 AM, M. F. Hasler <seqfan at hasler.fr> wrote:
>>
>> I think that such a taxonomy could be realized by allowing arbitrary custom
>> keywords
>> (maybe slightly moderated by Assoc.Editors), which would work like
>> mediawiki's "Categories",
>> and this could be an easy and efficient way to index, classify, and
>> cross-reference related sequences.
>> See oeis.org/wiki/User:M._F._Hasler/Categories for more details.
>>
>> Regards,
>> MH
>>
>> On Sun, Apr 8, 2018 at 6:13 PM, Wayne VanWeerthuizen wrote:
>>
>>> On 4/7/2018 1:17 PM, Georg.Fischer wrote:
>>>> But I think it is totally impossible to *manually* introduce any
>>> additional
>>>> classification of the OEIS sequences (because of their huge number).
>>>
>>> Again, I did not intend to suggest we classify every sequence. I was
>>> asking about a consistent nomenclature for simply talking about the
>>> different general ways sequences are generated. It's the list itself of
>>> possible categories that interests me, and not necessarily actually having
>>> every specific sequence categorized and tagged. And I'm not convinced that
>>> the huge number of sequences implies a huge number of meta-rules for
>>> sequence generation. I still expect a list of meta-rules to be relatively
>>> short, given the vast majority of OEIS sequences appear to be generated by
>>> essentially quite similar procedures. Again, one of my examples was that
>>> the OEIS already has hundreds, maybe even thousands of, "least k such
>>> that..." sequences. What other kinds of procedures are akin to this?
>>>
>>
>> --
>> Seqfan Mailing list - http://list.seqfan.eu/
> 
> 
> --
> Seqfan Mailing list - http://list.seqfan.eu/
>