LaTeX

[Geoffrey Critzer]

Greetings sequence fans

I have always wondered why the formulas in the data base are not type set in LaTeX.

Best regards

[Neil Sloane]

The reason is that the database is used by humans, not robots. Humans prefer to see formulas written the way you write them on a piece of paper.  Not disguised in a fog of \fracs, dollar signs, etc.

Best regards

Neil 

Neil J. A. Sloane, Chairman, OEIS Foundation.

Also Visiting Scientist, Math. Dept., Rutgers University, 

Email: njas...@gmail.com

--
You received this message because you are subscribed to the Google Groups "SeqFan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to seqfan+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/seqfan/60f6b76d-ac60-4dfc-a525-efb0be24cba9n%40googlegroups.com.

[Geoffrey Critzer]

Greetings and happy new year, Neil Sloane

I am sorry that I didn't make my question clear.

What I mean is why don't the formulas appear with the standard math symbols like we would see in a journal article or text book.  

For example, a capital sigma instead of Sum_{n>=1} ...

Best regards

Geoffrey Critzer

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAAOnSgQDeYojydsGR94i7mofWh7Nz6eU3M%2BcfE2a5OqqUss_fQ%40mail.gmail.com.

[Allan Wechsler]

The underlying data structure of OEIS is a "database" that is essentially pure formatted plaintext. The OEIS pages you see are database entries processed through an extremely simple "renderer", which basically just wraps boilerplate HTML around the different fields of a database record.

Geoffrey, if I understand your proposal, you want to gussy up the renderer for the FORMULA field, using (I imagine) some publicly available LaTeX program. Perhaps I don't have your proposal right, but if I do, it imposes two non-trivial costs on the maintainers of OEIS. The obvious cost is changing the code so that it delegates the FORMULA field to some off-the-shelf LaTeX renderer. But the much bigger cost is that all of almost four hundred thousand FORMULA sections will need to be converted in order to conform to LaTeX syntax.

It also imposes a new cost on OEIS contributors, namely, that they have to know LaTeX in order to produce a correct FORMULA section. And I guarantee that there are many contributors who don't know LaTeX. I am one of them. This is mere laziness on my part. I could learn LaTeX -- in fact, I knew it once, forty years ago. But it is a nonzero cost.

I suppose one could partly ameliorate these challenges by making LaTeX formatting optional, but that doesn't help a LaTeX-ignorant contributor who wants to add a new formula to an existing FORMULA section that is already LaTeX-enabled. Okay, you could make the LaTeX-enabling per formula by adding a new formatting convention. But this would be a lot more work for the maintainers: it would make the renderer for the FORMULA section at least five times as complicated as it is now. I think I would be in favor of such a change in principle, if it didn't require every contributor to learn LaTeX, and if it didn't make a lot more work for OEIS maintainers.

-- Allan

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CALmqthDaXDLGn7MLz%3D3uM9r_OyKML1gr_9XNdLZMSm5VuvUpUA%40mail.gmail.com.

[Sean A. Irvine]

Additional reasons:

* The OEIS predates Unicode and browser ability to display such characters.

* Such characters cannot easily be typed or edited (unless something like LaTeX is adopted for their representation).

* Such characters are actually less amenable to automated tools -- just look at how many potential Unicode characters could be used for multiplication. Even for the case of Sigma you would have to decide between U+03A3, U+2211, and various bold and italic forms which are also defined in Unicode.

As Allan indicates, in principle there is nothing that prevents you from developing your own rendering of the OEIS content.

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CALmqthDaXDLGn7MLz%3D3uM9r_OyKML1gr_9XNdLZMSm5VuvUpUA%40mail.gmail.com.

[Arthur O'Dwyer]

On Thu, Jan 1, 2026 at 2:07 PM Allan Wechsler <acw...@gmail.com> wrote:

The underlying data structure of OEIS is a "database" that is essentially pure formatted plaintext. The OEIS pages you see are database entries processed through an extremely simple "renderer", which basically just wraps boilerplate HTML around the different fields of a database record.

Geoffrey, if I understand your proposal, you want to gussy up the renderer for the FORMULA field, using (I imagine) some publicly available LaTeX program. Perhaps I don't have your proposal right, but if I do, it imposes two non-trivial costs on the maintainers of OEIS. The obvious cost is changing the code so that it delegates the FORMULA field to some off-the-shelf LaTeX renderer. But the much bigger cost is that all of almost four hundred thousand FORMULA sections will need to be converted in order to conform to LaTeX syntax.

It also imposes a new cost on OEIS contributors, namely, that they have to know LaTeX in order to produce a correct FORMULA section. And I guarantee that there are many contributors who don't know LaTeX. I am one of them. This is mere laziness on my part. I could learn LaTeX -- in fact, I knew it once, forty years ago. But it is a nonzero cost.

I suppose one could partly ameliorate these challenges by making LaTeX formatting optional, but that doesn't help a LaTeX-ignorant contributor who wants to add a new formula to an existing FORMULA section that is already LaTeX-enabled. Okay, you could make the LaTeX-enabling per formula by adding a new formatting convention. But this would be a lot more work for the maintainers: it would make the renderer for the FORMULA section at least five times as complicated as it is now. I think I would be in favor of such a change in principle, if it didn't require every contributor to learn LaTeX, and if it didn't make a lot more work for OEIS maintainers.

At first glance it seems "easy enough" to make a rule in the spirit of Markdown: basically whenever there's a string bracketed by dollar signs, a renderer should be able to assume the stuff inside the dollar signs is LaTeX math syntax. OEIS contributors should please be sensible enough not to use $this syntax$ for anything they don't want rendered as TeX.

However, there are at least three caveats to this modest proposal:

(1) The actual "official" OEIS site will probably never render LaTeX, as Allan said, because that's work and fragile and not fun. Besides, your target audience can mostly read LaTeX just as easily as plain text — I mean, $\sum_{n\geq 1}n$ isn't much less readable than Sum_{n>=1}(n) — so little benefit accrues to "tangling" that LaTeX syntax into rendered mathstuff.

(2) I've done this kind of thing on my blog, and occasionally found this annoying problem: The stuff you put in $...$ is just a little tiny piece of a proper LaTeX document. Is it safe to use things like \cfrac and \dotsb inside those expressions? Well, that depends on whether your LaTeX document started by importing the `amsmath` package or not. You want to be able to tell the OEIS editor, "Just use LaTeX syntax and it'll work"; but to a first approximation there's no such thing as "LaTeX syntax" — the syntax is entirely determined by what packages you're using. Even so-called "plain TeX" is mostly library code; very little functionality is built into TeX-the-program by itself. So at the least you'd need a wiki page detailing which packages the renderer imports by default, and an editorial process for requesting the enablement of new packages (or the upgrading of old ones). Again, that's a lot of work, and fragile and unfun. (Wikipedia must have such a process somewhere, although I looked and couldn't find it. See e.g. here.)

(3) We like to pretend that TeX syntax is "just for math," but again, it's a system built for entire documents. The "bracket things in dollar signs" convention itself is part of TeX syntax. Should there be a way for OEIS editors to write non-math using TeX syntax too? Almost certainly not, but it's a can of worms that would demand at least a quick peek inside, if OEIS were to start a project to bless "TeX syntax." I have no good positive examples here. But for example I'm sure we do not want people to start submitting entries that say Poincar\'e (the TeX-markup syntax for an acute accent) in place of Poincaré (the plain-text Unicode representation of é).

Happy New Year,

–Arthur

[David desJardins]

It would be easy to build an AI-based frontend that translates whatever is in OEIS into formatted LaTeX output.

--

You received this message because you are subscribed to the Google Groups "SeqFan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to seqfan+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CADvuK0L-BouOrGKq00W0DZq3NfvXbzOAZ2Fyp7Vi9%3DGJL3vC-w%40mail.gmail.com.

[Jonas Karlsson]

The other answers make good points, but they also seem to assume that displaying symbols would be a good thing if it could be done simply, reliably etc. I, for one, prefer plain text, and I wish more websites would stick to it.

J

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAP%3DxTqN9CxLgzvCrnHB2k9HgcEn-zGpLnicbzAznMUE2dryaQQ%40mail.gmail.com.

[Ed Pegg]

I asked ChapGPT for https://oeis.org/A000108 to be rendered. 
It's easier to leave OEIS as it is now.  

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CABiOuo_Uzg9wurMj4WEWAmWC3gCY9ZeT-grPcVv00QQzSfK0Eg%40mail.gmail.com.

[David desJardins]

On Thu, Jan 1, 2026 at 1:54 PM Jonas Karlsson <jonas...@gmail.com> wrote:

The other answers make good points, but they also seem to assume that displaying symbols would be a good thing if it could be done simply, reliably etc. I, for one, prefer plain text, and I wish more websites would stick to it.

Giving people the option seems like always a good thing. 

[Andrei Zabolotskii]

As a human, I certainly prefer the formulas to be displayed the way I write them on a piece of paper. This is currently not how it works in the OEIS, unfortunately.

In fact, the problem of inserting inline formulas into texts posted online has been solved long ago: Wikipedia, Mathstodon, and databases like LMFDB display formulas properly (like in the printed EIS!).

Please put your "thumbs up" on this OEIS bug tracker issue if you agree that we need this: https://github.com/oeis/oeis/issues/28 "Proper rendering of formulas"

Andrei

четверг, 1 января 2026 г. в 22:09:42 UTC, edp...@gmail.com:

[David desJardins]

I don't support "Proper rendering of formulas #28" as phrased. It doesn't seem to be a request about rendering, but a request about input. As stated above, it's easy for modern AI systems to take random human-generated text and produce nicely rendered LaTeX style output, if that's what you want. We aren't living in the 1990s. So if what you want is rendering, then trying to get thousands of contributors to write LaTeX just is not the right solution, in my view. There's no need for that.

To view this discussion visit https://groups.google.com/d/msgid/seqfan/d4ff44a0-34d9-448b-ad4e-abae8b3234f1n%40googlegroups.com.

[Andrei Zabolotskii]

What I am asking for is an option to input a formula in LaTeX (or in any other suitable way) and get it rendered in the OEIS (note that I wrote: "We need to be able to..."). This is how it already works in e.g. Mathstodon (I am deliberately giving an example of an independent website used by professional mathematicians and maintained by just 2 people).

You received this message because you are subscribed to a topic in the Google Groups "SeqFan" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/seqfan/rmxUA-JwUcA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to seqfan+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAP%3DxTqO%2B2evEn99_a8TX41xTUKXqSV8HKn2NatPi_60oBtYdEg%40mail.gmail.com.

[David desJardins]

On Thu, Jan 1, 2026 at 3:40 PM Andrei Zabolotskii <andrei.d.z...@gmail.com> wrote:

What I am asking for is an option to input a formula in LaTeX (or in any other suitable way) and get it rendered in the OEIS (note that I wrote: "We need to be able to..."). This is how it already works in e.g. Mathstodon (I am deliberately giving an example of an independent website used by professional mathematicians and maintained by just 2 people).

Sure. I'm not saying you can't request that. I'm just saying (1) I don't think that's particularly important, and (2) if what you want is 'proper rendering' of math formulas, that's not the best way to get it.

[Victor Miller]

If the web page uses MathJax you can use AsciiMath 

https://asciimath.org/

--

You received this message because you are subscribed to the Google Groups "SeqFan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to seqfan+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAP%3DxTqOwPgzPgq7DC0%3DLAgasjFuNr2n3LzeGT1fpdiu1O%2Ben_g%40mail.gmail.com.

[Antti Karttunen]

The problem with (the standard variants of) LaTeX is that it still would not allow using most of the vast repertoire of Unicode characters. And in any case, the traditional mathematical notation mostly sucks. Like for example, mathematicians are still using the zillion times overloaded set of standard Greek letters for all the function names (we got Carmichael's lambda and Liouville's lambda, Ramanujan's tau and someone else's tau, and so on), and as far as I know, LaTeX doesn't even allow using such less known Greek letters as ϡ (Sampi), Ϝ (Digamma), Ϙ (Koppa), or Ͱ (Heta). Too bad for a glyphophile like me!

But never mind and happy ፳፻፳፮ ! (See https://oeis.org/A098378 Although I recall that they count their years differently there in Abyssinia).

- Antti

[Antti Karttunen]

(Changed the subject, as I hijacked this thread for my personal fix-idés).

On Sat, Jan 3, 2026 at 8:17 AM Antti Karttunen <antti.k...@gmail.com> wrote:

The problem with (the standard variants of) LaTeX is that it still would not allow using most of the vast repertoire of Unicode characters. And in any case, the traditional mathematical notation mostly sucks. Like for example, mathematicians are still using the zillion times overloaded set of standard Greek letters for all the function names (we got Carmichael's lambda and Liouville's lambda, Ramanujan's tau and someone else's tau, and so on), and as far as I know, LaTeX doesn't even allow using such less known Greek letters as ϡ (Sampi), Ϝ (Digamma), Ϙ (Koppa), or Ͱ (Heta). Too bad for a glyphophile like me!

But never mind and happy ፳፻፳፮ ! (See https://oeis.org/A098378 Although I recall that they count their years differently there in Abyssinia).

It is interesting to note that with a similar Greek numerals system, this is expressed differently, as ͵ΒΙΚϚ

And if I give that string to Google AI, it answers in Greek (with some Finnish as well):

---

Mietitään

Ο αριθμός

͵ΒΙΚϚ στο ελληνικό σύστημα αρίθμησης αντιστοιχεί στο έτος 2026.

• ͵Β = 2.000
• Ι = 10
• Κ = 20 (Σημείωση: Συνήθως χρησιμοποιείται το Κ για το 20 και το Ϛ για το 6)
• Ϛ (στίγμα) = 6

• 
  

Σήμερα είναι Σάββατο, 3 Ιανουαρίου 2026.

Tekoälyn antamat vastaukset voivat sisältää virheitä

---

Then..., because somebody could interpret the Ethiopic numeral string naively as ፳፻፳፮ as ፳ = 2, ፻ = 0, ፳ = 2, ፮ = 6, although it really is ፳ = 20, ፻ = 100, ፳ = 20, ፮ = 6, I got a vague inspiration for the sequence of numbers where such an ambivalence exists. Or something more generic.

Funny that in so few places they made the final leap of inventing 0 and a true positional system, although already the Greek numeral system was already quite near. And then the Roman numerals was just a big step backwards.

Best regards,

Antti

[Antti Karttunen]

On Sat, Jan 3, 2026 at 9:19 AM Antti Karttunen <antti.k...@gmail.com> wrote:

(Changed the subject, as I hijacked this thread for my personal fix-idés).

On Sat, Jan 3, 2026 at 8:17 AM Antti Karttunen <antti.k...@gmail.com> wrote:

The problem with (the standard variants of) LaTeX is that it still would not allow using most of the vast repertoire of Unicode characters. And in any case, the traditional mathematical notation mostly sucks. Like for example, mathematicians are still using the zillion times overloaded set of standard Greek letters for all the function names (we got Carmichael's lambda and Liouville's lambda, Ramanujan's tau and someone else's tau, and so on), and as far as I know, LaTeX doesn't even allow using such less known Greek letters as ϡ (Sampi), Ϝ (Digamma), Ϙ (Koppa), or Ͱ (Heta). Too bad for a glyphophile like me!

But never mind and happy ፳፻፳፮ ! (See https://oeis.org/A098378 Although I recall that they count their years differently there in Abyssinia).

It is interesting to note that with a similar Greek numerals system, this is expressed differently, as ͵ΒΙΚϚ

And if I give that string to Google AI, it answers in Greek (with some Finnish as well):

---

Mietitään

Ο αριθμός

͵ΒΙΚϚ στο ελληνικό σύστημα αρίθμησης αντιστοιχεί στο έτος 2026.

• ͵Β = 2.000
• Ι = 10
• Κ = 20 (Σημείωση: Συνήθως χρησιμοποιείται το Κ για το 20 και το Ϛ για το 6)
• Ϛ (στίγμα) = 6

• 
  

Σήμερα είναι Σάββατο, 3 Ιανουαρίου 2026.

Tekoälyn antamat vastaukset voivat sisältää virheitä

Indeed... the iota (= 10) should have not been present! (But still the AI claimed that it would be equal to 2026, probably because that number is trending just now?)

So, without superfluous numeric letters, I got just this:

Ο αριθμός

͵ΒΚϚ στο ελληνικό σύστημα αρίθμησης αντιστοιχεί στο έτος 2026.

• ͵Β = 2.000
• Κ = 20
• Ϛ = 6

Tekoälyn antamat vastaukset voivat sisältää virheitä.

Best,

Antti

[M F Hasler]

On Thu, Jan 1, 2026 at 3:20 PM Sean A. Irvine <sai...@gmail.com> wrote:

* The OEIS predates Unicode and browser ability to display such characters.

That's not true: Unicode was designed in the 80s but became a standard in 1991, 

while the OEIS was launched in 1996 ;

by that time most browsers did support unicode.

(To get an idea, Apple decided to incorporate Unicode support into TrueType in April 1988.)

 

* Such characters cannot easily be typed or edited (unless something like LaTeX is adopted for their representation).

there are three main approaches: \LaTeX format (as in Mediawiki's <math> tags),

HTML entities (e.g., &ge; for >= ; &sum; for the summation symbol, &pi; etc. for greek letters, ...)

and direct "typing" using an extended keyboard, often realized as a "symbol bar" above or below the text area,

see for example https://en.wikipedia.org/w/index.php?title=OEIS&action=edit

or this screenshot if it comes through:

I think the current policy of allowing Unicode only in foreign names but nowhere else is obsolete.

If we'd *allow* to use the more human and less "robotic" notation  x² , π , ... 

instead of x^2, Pi, ... (unquestionably artifacts of computer programming  languages)

at least in comments (but I'd suggest: everywhere),

that doesn't mean in any way that we *enforce* it , nor that we have to convert all existing material to Unicode!

The "searchable" argument also isn't one, since math symbols and tiny words like "in", ... are anyways stripped from search.

BTW, it's not difficult to make the search engine recognize  π  and  Pi  as equivalent,

in exactly the same way as a search for "Erdös" also finds "Erdos" and vice versa.

This is already built in into all multilingual text handling libraries, in particular those used in OEIS software:

For example, try typing "Erdös"  into the OEIS search box, it will say that you are searching for "erdos",

and the first hits will be sequences with "Erdős" highlighted in their NAME

(with the correct Hungarian " instead of the ¨ I used).

Actually, it will most likely even improve search results if ≥ and \ge and &ge; are all recognised as equivalent

(just like ö = ő = o),

rather than the current situation where >= is completely ignored (if I'm not wrong).

-Maximilian

On Fri, 2 Jan 2026 at 05:45, Geoffrey Critzer <geocri...@gmail.com> wrote:

What I mean is why don't the formulas appear with the standard math symbols like we would see in a journal article or text book.  

For example, a capital sigma instead of Sum_{n>=1} ...

 

On Thu, Jan 1, 2026 at 10:27 AM Neil Sloane <njas...@gmail.com> wrote:

The reason is that the database is used by humans, not robots. 

Humans prefer to see formulas written the way you write them on a piece of paper. 

 

exactly -- that is:  x²  and  𝜋  rather than  x^2  and  Pi !

We could still *enter* them as x^2 and &pi; or \pi, but have them displayed as x² and 𝜋.

[David desJardins]

My view would be that using Unicode characters like "superscript 2" is a pretty inferior stopgap to almost any alternative. It doesn't really generalize to arbitrary mathematical notation. There's a fundamental limitation with assembling math formulas this way.

That's kind of different from Unicode "pi", which is reasonable enough if you want to use it.

--

You received this message because you are subscribed to the Google Groups "SeqFan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to seqfan+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAFqvfd8bcPf45b_XK-wTCJndDe4agA%3DjWB7OQ9gO9q6xKkxYkg%40mail.gmail.com.

[Dave Consiglio]

My two cents: I often find the non-Latex format so difficult to read that it can significantly hinder my ability to understand what's being described. It seems to me that keeping what's there and adding Latex (perhaps as an additional field) would mitigate any fears of loss of information while greatly improving readability. 

To view this discussion visit https://groups.google.com/d/msgid/seqfan/CAP%3DxTqO_qfnoB2vzFjZsWq9khTaK1ovj3Oh%2BFhPVeJUS-q2G4g%40mail.gmail.com.

[jp allouche math]

My two-cents: I feel like this discussion overlaps somewhow a several 
decade old debate between wysiwyg and non-wysiwyg fans (debate to 
which there is---afaict---no universally accepted solution).
best
jean-paul

[David desJardins]

On Mon, Jan 5, 2026 at 11:40 AM Dave Consiglio <dave...@gmail.com> wrote:

My two cents: I often find the non-Latex format so difficult to read that it can significantly hinder my ability to understand what's being described. It seems to me that keeping what's there and adding Latex (perhaps as an additional field) would mitigate any fears of loss of information while greatly improving readability.

Your problem makes perfect sense, but it just seems like the wrong solution. Now that we have software tools that are completely capable of parsing what we currently have and showing you what you want to see, why don't we just use those?