Given that most people’s reaction to the question “should BHL charge for access?” is likely to be an emphatic NO! let me explain my goal here. I have no formal role in BHL, this post is purely me shooting the breeze, and hoping that BHL is looking at all the options as it considers its future.
The sustainability of BHL is in doubt, the current library membership model relies on funding from organisations that are facing their own budgetary constraints, and there does not seem to be an obvious source of the kind of funding that would help BHL grow, yet alone survive. BHL needs money. Yet, charging for access seems utterly opposed to its goal of free and open access to biodiversity literature. Can we reconcile these two things?
Consider the academic search engine OpenAlex. In answer to the question “If OpenAlex is free, why are you charging money?”, their FAQ states:
Our data is open and free: everything is released under a CC0 public-domain license, with no “personal use only” carve-out and no permission to ask. You can even download the entire dataset for free. However, the services on top of the data — the API, search, sync, and support — are expensive for us to provide, and so we charge for those services in order to keep OpenAlex sustainably open.
In justifying their pricing OpenAlex cite " the POSI principle of revenue based on services, not data.".
OpenAlex do a clever thing with their API pricing: you get $1 per day free. So most casual users will not be charged, the target, ironically, is people like me who routinely pull lots of data from APIs.
The emphasis on paying for services not content means that you are not charging for open access content, you are charging for the ways you enable that content to be discovered and used.
I think BHL has an opportunity to consider this route because the next generation of digital libraries are going to be expensive. BHL needs to re OCR the entire corpus. I asked Claude to price that, and state of the art OCR such as Mistral 4 would cost $US 128,000 for 64 million pages. There are older models that could do it for less (Mistral 3 is $US 64,000), but the point is, that is a significant cost. The kind of image search that I have been experimenting with on a small scale (see BHL image search) would require significant investment in a large server to host the embeddings for those images, and that is before we factor in indexing the likely millions of individual figures that we will be able to extract using state of the art OCR. Likewise I would argue that the current fulltext index of BHL needs to be redone because it consistently yields counter intuitive results. And all of these tools will need to be continuously updated as BHL grows. Note also that a lot of BHL runs on services that are “free” through the good graces of organisations such as Internet Archive and the AWS Open Data program. If either of those services stopped being free then BHL would have a significant problem.
Given that BHL is also redesigning the platform, there is an opportunity to rethink the notion of “free”. For example, one approach would be:
- keep the current BHL interface as “classic” BHL, which is free
- create new portal that has basic free level (e.g., looking at content)
- users wanting more sophisticated access need to create an account (I know, I know, super annoying), they get a free quota
- users spending a lot of time on the site, using the API a lot (and/or using services that are expensive to provide) pay for that privilege.
Would this work? Are there enough people or organisations wanting these services? What added value can BHL offer? Will the free tier be enough to maintain the community of enthusiastic but low volume BHL users from across the academic, artistic, and other communities? I don’t know. But I think there’s a conversation to be had.