I’m giving up writing at Medium

Standard

I’ve been ser­i­ously think­ing about mov­ing to Ghost or Medium for writ­ing. Ghost uses Markdown, which I like is handy for when I write in StackEdit​.io. Medium has a very simple inter­face. It’s not cus­tom­is­able, but the flip side of that is that you don’t waste time try­ing to cus­tom­ise it.

I gave Medium a go with two short stor­ies I’d writ­ten. I was going to post some ser­i­ous and researched sci­ence art­icles, but I chose to put up the short stor­ies as I’d not be bothered if got no views. Here are the res­ults.
Continue read­ing

How did being buried for 36 hours become three days?

Standard

Something that puzzled me about the resur­rec­tion was how a period of thirty-six hours or so became three days. There are other things too, but the period from death to Easter morn­ing isn’t even forty-eight hours. Where does three days come from? Couldn’t ancient people count?

It turns out they could, but they coun­ted differently.

Possibly praying that he doesn't have to sort out the numerical problems.

Possibly pray­ing that he doesn’t have to sort out the numer­ical problems.

In ancient Greece and Rome they used inclus­ive count­ing. This is where you count the first and last things in a series. For example, how often are the Olympics held? We would say every four years. The Greeks would have said every five years, and they called it a pen­teric fest­ival. Here’s how you get five years for the Olympics.

Year one: Hold the Olympics.
Year two: The Isthmian and Nemean Games.
Year three: The Delphic Games.
Year four: The Isthmian and Nemean Games (again).
Year five: The Olympic Games.

The Romans also used this sys­tem of inclus­ive num­ber­ing for their cal­en­dar. Jerusalem at the time was in the Roman Empire.

Counting of the days where you start and fin­ish is what gives three days. Jesus has to die before sun­set on the Friday. The reason for this is at sun­set a new day starts in the Jewish cal­en­dar. This second day car­ries on to sun­set on what we could call Saturday. At sun­set the third day starts. Now Jesus can rise any time he likes and he’ll have risen on the third day.

With care­ful tim­ing he could have kept it down to just over twenty-four hours.

Whether or not it happened is another dis­cus­sion, but inclus­ive count­ing shows why the ancients were happy to say ‘on the third day’, even though they knew it was well under two full days.

Edit: Bill Thayer has more fest­ivals with inclus­ive count­ing.

5 Years On — Chemotherapy Works

Standard

I wrote someone out of my will today.

It was five years ago I had chemo­ther­apy for can­cer. It should have been six, but I held off get­ting a dia­gnosis because I was in the last year of my PhD and help­ing out with eld­erly rel­at­ives. I wasn’t strictly in denial about hav­ing can­cer, but the tim­ing was bad. Relatives died which caused more prob­lems. When another close rel­at­ive was hos­pit­al­ised it was obvi­ous there wasn’t going to be a con­veni­ent time.

I was dia­gnosed on a Monday after­noon and oper­ated on the fol­low­ing day. It wasn’t that bad a situ­ation, someone else had can­celled their oper­a­tion due to snow. I was offered either their spot, or else wait a few weeks for the oper­a­tion. Hanging around with the tumour inside me seemed like a really bad idea, so in I went. The follow-up was a brief course of chemotherapy.

There’s been a lot writ­ten about how bad chemo­ther­apy is, but I had no prob­lem. Here’s a selfie from five years ago while I’m hav­ing chemotherapy.

chemo selfie

I pottered around the house and had no trouble at all. I didn’t have any prob­lem, though one day I did fancy some Jaffa Cakes and there were none in the house. So I went out to the shops to get some. This is a map of how far away the shop was.

Map via Google Maps.

Map via Google Maps

I was tired well before the first corner. Continue read­ing

A New #WordPress plugin for #SCIENCE!

Standard

…and other research too.

I have a work­ing first release of a plu­gin to link to research in a ScienceSeeker friendly way in a WordPress blog. It will only work with self-hosted WordPress installs, it will not work on WordPress​.com blogs.

The way it works is you enter the ID(s) of the thing(s) you want to include then, when you save the draft, the web­site pro­duces a format­ted cita­tion that it will auto­mat­ic­ally append to the con­tent of your post. It will also add a META tag to the head of the page. This will give a way to tell sites like alt​met​ric​.com what paper(s) your blog post is about.

Editing in WordPress

Editing Screen. Click to embiggen.

Citation output

The out­put. Click to largify.


It’s primar­ily built to work with DOIs, because that’s what we use most at AoB Blog. You can type in a DOI as 10.1093/aob/mct168 or http://​dx​.doi​.org/​1​0​.​1​0​9​3​/​a​o​b​/​m​c​t​168 and when the site saves it will get bulked out to the longer ver­sion. You can enter more than one entry, so stick­ing 10.1093/aob/mcp121, 10.1093/aob/mcs287, 10.1093/aob/mcq238 should work too.

Once the ref­er­ence is parsed, it appears as a cita­tion. When you have this cita­tion, you can edit it in this box. You might need to do that if the pars­ing breaks. It’s pos­sible some DOIs will give inform­a­tion in a dif­fer­ent way to most. Currently the plu­gin works with stand­ard DOIs and figshare’s DOIs. It’s very pos­sible there are some other sites that have their own stand­ards so, if you find one, let me know.

To clear the ref­er­ences and cita­tions on a post, delete all the ref­er­ences and save. The plu­gin will wipe the cita­tion box.

You can add arXiv ref­er­ences. I’ve set it so you copy and paste a URL from arXiv to the box to give http://​arxiv​.org/​a​b​s​/​1​3​0​6​.​5​148. If there’s demand it should be pos­sible to send any nine char­ac­ter ref­er­ence with a dot in the middle to the arXiv mod­ule. I’ve spot­ted a bug in the arXiv mod­ule put­ting together the screen shotes (look at the author name). I think I’ve fixed this.

It’s not so good for the Social Sciences and Humanities. Here mono­graphs are still import­ant research out­puts, which means ISBNs. These are more of a prob­lem. You enter them as a straight run of ten or thir­teen char­ac­ters. The only place I’ve found giv­ing inform­a­tion from ISBNs in a friendly format is Google Books. But from here I can only get Title, Authors and Publisher. I can­not get Publisher Location from the data.

For DOIs and arXiv papers it’s obvi­ous to link through to the paper. Books tend not to have a recog­nis­able home page. I’ve linked through to Google Books because that’s where the data comes from. But it’s pos­sible that LibraryThing or the Amazons would be bet­ter places to link to.

This sys­tem doesn’t handle book chapters yet, unless they have a DOI. Lying in bed I thought it could be handled as Chapter Authors::Chapter Title::Page Start::Page End::ISBN and any­thing with a double colon gets passed to a book chapter mod­ule for format­ting. I’m not sure if this is use­ful, or if it’s get­ting to stage where typ­ing the ref­er­ence in is more effort than it’s worth.

At the moment the link is on the iden­ti­fier, because that’s the way Research Blogging and ScienceSeeker work. Alan Cann has sug­ges­ted mak­ing the whole ref­er­ence click­able. I’m not sure if this is a good idea or not. It’s a big­ger click­able tar­get, and CSS styl­ing makes the present­a­tion a mat­ter for whoever’s site it is.

The plu­gin doesn’t work for Research Blogging yet. Research Blogging needs ref­er­ences asso­ci­ated with a sub­ject. The first way I’d writ­ten this meant that sub­jects would have to be hard­wired in. Now I think it should be pos­sible to tweak the plu­gin to add Research Blogging top­ics on a post-by-post basis, but not (yet) on a citation-by-citation basis. This would work for most people cit­ing just one paper in Research Blogging posts, but some people cite mul­tiple papers in one post. The way I’m think­ing would label all cita­tions in one post as being the same topic.

Finally, like me, it doesn’t fail grace­fully. I’ve spent quite a while get­ting the damn thing to work. Deliberately break­ing it, so I can make it fail nicely, hasn’t enthused me yet.

You can down­load it from my Dropbox at https://​www​.drop​box​.com/​s​/​k​b​0​w​0​2​j​r​3​4​a​g​r​2​v​/​r​e​s​e​a​r​c​h​l​i​n​k​s​.​zip. You install it by going to your plu­gin menu and upload­ing the zip file. You make sure you upload it to your test site, because this is still beta soft­ware. I think this will be com­pat­ible with the final ver­sion, but I’m not will­ing to guar­an­tee. If you have installed the pre­vi­ous ver­sion, this ver­sion is utterly incom­pat­ible and using the two at the same time will break access to your blog in a very emphatic way. This is why I test on a desktop server.

I’ll be test­ing this shortly, in par­tic­u­lar the way it handles COinS. There may be a simple and eleg­ant way of adding COinS to ref­er­ences, but I don’t know what it is.

WordPress shortcodes for DOIs and other research links

Standard

I’ve writ­ten a plu­gin for WordPress that adds short­codes to link to research. To use it you’ll need a self-installed WordPress blog. A blog at word​press​.com won’t be able to use it.

[doi id=“10.1093/aob/mct148”]link text[/doi] links the link text to http://​dx​.doi​.org/​1​0​.​1​0​9​3​/​a​o​b​/​m​c​t​148 It also does a few more things.

If there are research codes used, the blog checks to see if it has cita­tions stored for them. If it does then it out­puts a References sec­tion at the end of the post. It’s not in this post, because I’m not actu­ally using the short­code in this post.

If there isn’t a cita­tion stored, the blog will go through each DOI and make one. It’s designed to make cita­tions com­pat­ible with the ScienceSeeker web­site. If you’re a sci­ence blog­ger then you would do well to sign up to this.

It solves a prob­lem with alt­met­rics too.

At AoB Blog we post about vari­ous papers, usu­ally from Annals of Botany or AoB PLANTS. If people like what they see and tweet a link to the post, the post gets credit for the link, but the ori­ginal paper doesn’t. It wasn’t the ori­ginal paper tweeted, it was the post. I’ve talked with people at OUP and Altmetric​.com and the answer we settled on was to add a meta tag in the header.

<meta name=“DCTERMS.isBasedOn” scheme=“DCTERMS.URI” content=“http://dx.doi.org/10.1093/aob/mct148”/>

This gives a way to con­nect a post to the paper it’s com­ment­ing on. It only works from the second time a page is loaded, but the first is usu­ally a pre­view by the author, so that’s not critical.

Not everything people want to link to has a DOI. What do you do if you have a mix of DOI and http:// links and you want to use them in the same ref­er­ence sec­tion? There is a url shortcode.

[url id=“http://aobblog.com/2013/08/brachypodium-is-not-arabidopsis/” author=“Chaffey, N.” date=“2013” title=“Brachypodium is NOT Arabidopsis(!)” website=“AoB Blog”]

…will out­put a cita­tion in a sim­ilar style to the DOI cita­tion, with a ‘date accessed’ note added, set to whenever you first run the short­code. If you don’t spe­cify a pub­lic­a­tion date then it sets this year. If you don’t spe­cify an author then it is set to ‘Unknown’.

You can down­load the code from Dropbox and install it, or have a laugh the at code. The code is not sleek, because I wanted to see what was hap­pen­ing each step of the way. As a warn­ing, I don’t know if this is secure code or not. That’s not likely to be a big prob­lem if it’s just you on your site, but it’s an issue if you have a multi-author blog.

Also it doesn’t handle a few other things yet. There is a space for an [arxiv] short­code. I haven’t added this yet because arxiv out­puts metadata in a dif­fer­ent way to dx​.doi​.org. Despite hav­ing DOIs, Figshare doesn’t work with it either. I don’t know why Figshare out their DOIs in a dif­fer­ent way, because I haven’t spoken to them yet. There’s prob­ably a good reason, so that might mean mak­ing a [fig­share] short­code to handle those links.

At the moment the code is up for dis­cus­sion. Once I’ve under­stood this page and added arxiv and fig­share sup­port then I’ll see about adding it to the WordPress plu­gin repository.

Download the plu­gin as a zip file

Update: Thanks to Stack Overflow I now know how to get data for an ISBN, so an [isbn] short­code will be pos­sible too.