For any project like this, web hosting for MediaWiki, and funding to pay for hosting, are related issues. I’m still experimenting so I’m not sure how powerful a server the project will need in the long term or how much money I’ll be able to raise.

Hosting

The wiki is currently hosted by Krystal. I’ve used them for other things so I know they’re very good. For £12 a month, I get cloud hosting on a Litespeed server with:

• 1 CPU core
• 2GB RAM
• 50 entry processes
• ‘unlimited’ disk space, bandwidth and SQL databases
• daily backups
• Cpanel
• free SSL certificate
• SSH access (essential for installing and maintaining Semantic MediaWiki)
• Composer already installed globally (makes installing and upgrading Semantic MediaWiki much easier)

My first impression is that the wiki is running really fast. I don’t know how long this will last as the number of pages and number of users increase. Once I’ve finalised the data structures and start importing batches of data, the wiki could easily grow to tens of thousands of pages (bigger than Marine Lives) fairly soon. As long as I’m the only editor, everyone else is likely to get pages from the Litespeed cache, which takes a load off the server, but sooner or later I want to invite more people to register and edit, which might put more strain on the server.

Funding

I can cover £12 per month myself indefinitely, especially during the phase when the wiki is primarily for my own use, but it would be nice if I could raise money to cover it. If the project gets to the point where the current server isn’t enough, the cost of hosting and server management would go up drastically, so I would need to raise more money. It would also be nice if I could eventually get money to pay me for my labour and responsibilities, especially once I open up the wiki to other contributors, because then I would have to spend more time on support and moderation.

These are the fundraising methods that I’m thinking about:

• first of all, I want to see how much money, if any, I can get from referrals to online retailers (I’m not allowed to be any more specific than that). Because of this, importing modern books and linking them to their subjects will be a high priority once the data structures are finalised.
• the next thing to try will be inviting donations through Paypal. These can be one-off or a recurring subscription. There will be no backer rewards, stretch goals or other incentives. The donations will just be tips for something that I was doing anyway. This ensures that they will be completely outside the scope of VAT (one of the reasons why I wouldn’t use Patreon is that they automatically charge VAT, which increases the cost to backers). I don’t want to start asking for money like this until I’ve imported enough data to actually be useful for people’s research, which will probably take a few months.
• further in the future, once I’ve imported existing reusable datasets and my own research notes, I might try crowdfunding campaigns to pay me for specific tasks. One big task would be correcting the OCR text of the Internet Archive’s scans of Calendar of State Papers Domestic, doing record linkage on named entities, and importing it all into the wiki. For this kind of campaign I would use Kickstarter or something similar, although I’m not sure what kind of rewards I could offer.
• there’s a possibility of getting academic research funding to help with this project, but it’s a very small possibility. Some of the biggest sources of funding (AHRC, ESRC) aren’t open to me because I’m an independent researcher, so that would depend on somehow forging links with academic institutions that are eligible. The grants that are open to independents can be used for creating digital resources, but there are limits. British Academy Small Research Grants have to be for a self-contained project and can’t be put towards an existing bigger project. The Leverhulme Trust prefers not to fund projects that are primarily creating data rather than answering specific research questions. The way to go would probably be to define a project with more specific research questions and traditional outputs, then publish the data at By The Sword Linked as a side-effect. Even then, the chances of a successful funding bid are low for everyone, and independents can’t usually get paid for our time unless we’re hired as research assistants or transcribers on someone else’s project.

I’d be interested to know whether anyone has any thoughts on the issues raised here. How much server resources does a big Semantic MediaWiki need? Are there better hosting options? Are there any sources of funding I haven’t considered? Would you be willing to contribute to crowdfunding once the wiki becomes useful?