Is there a need for a logo? I appreciate John’s pragmatism, as often seen — this is not flattery. I, as a starter, plunged into the logo design topic forgetting, perhaps, its very purpose. Let's take a moment to reflect on that.

A logo is not here to provide graphical components and a favicon for websites and other documentation; though obviously it can be used for that.

The primary purpose of a logo is to uniquely identify “at first glance” the brand it represents. Its prime characteristics are that it must be succinct and efficient.

Now consider the world “pandoc”. It is succinct (2 syllables) and efficient (the word says it all about the software).

In effect “Pandoc” has become a brand of a widely used tool which sets the standard of the software niche it has created.

Pandoc adheres to the UNIX  philosophy: do one thing and do it well. This further enhances the quality — one could say the “robustness”, of the brand.

Don’t get me wrong here. Pandoc is a complex beast with a myriad capabilities. But from a user perspective it is “dead simple”. And that is the distinguishing factor of the “Pandoc” brand.

Pandoc, the  brand, is well established in the community and has been for over a decade now.

So, what distinguishing factor would a graphical component -- i.e. a logo, add to the brand ? Well I must admit after long meditation on the topic: none.

Consider the logos of companies like Coca Cola, Ford, IBM, Oracle. The logo is the brand’s name. The brand name (and its colour) allow you to immediately identify that brand.

Adding a graphical component to the “Pandoc” brand will not significantly enhance its visibility, will not make it gain more “market share”, will not further extend/enhance the “core business” of the Pandoc software, …  (because that is what PR is about).

And speaking of PR, one could even argue that a logo is not a good idea. In an all visual world, often with little content, "good old Pandoc, says little but gets the job done". And to be a little sarcastic, consider those legions of Javascript frameworks that have appeared in this decade and almost immediately disappeared... they had wonderful and ever so promising websites, logos, etc. whilst their implementation remained beta or even pre-beta.... where are they now?

Does that mean that all the discussions made in this thread should be trashed? No. Imagery to document Pandoc on various media is necessary. Obviously. Be it in print or digital.

Such imagery is necessary to create consistency and further enforce the brand’s “robustness”.  A favicon, a graphical representation to beautify documents and blogs, a symbol to use in workflows and engineering diagrams, etc.. all are needed. But they should probably not be called a logo, least “the Pandoc logo”.  They should simply be part of a standard visual statement designed specifically for Pandoc.

(Remember that the colour codes are as important — if not more, than the graphical representation. All components of the visual statement must adhere to the “official” colour set. Think of Coca Cola’s red or McDonalds yellow :-)

As food for thought, consider AWS (Amazon web services). The words "AWS" and "Amazon", like "Pandoc", reflect the quality and value of the associated products. Amazon has created (and actively maintains) a unique and distinguishing stencil set for engineering purposes (https://aws.amazon.com/architecture/icons/). This is, in my (modified) opinion, what this thread should about. Just strip the word "logo" and keep going :-)

Thanks John.

Cheers

Jean-Michel

PS: I initially started this reply as a long discussion on the pros (and cons) of one letter logos — the most famous being McDonalds’ M — which for memory doesn’t (or didn’t) represent an M but two (initially overlapping) arches of their first (high tech) franchise. I stripped it because that would have ended up as a long (and possibly boring) dissertation. In essence none of the pros really apply to software products in general, and Pandoc in particular. Happy to drill this down if needed.