hit counter

I'm currently working on a DPhil in HCT at the University of Sussex. This section of the website is for an on-going 'learning diary', for me to write my thoughts and notes on various courses and my thesis.

Entries in process (8)


Thesis tech 2 - BibTeX

I'm already pretty comfortable with the workflow I outlined in Thesis tech, so the next step was looking at how to reference. LaTeX and BibTeX obviously play well together, but I've been using Mendeley to store all of my references. At this stage I don't really want to swap and use BibTeX, plus the import features on Mendeley are really handy. 

Fortunately for me, Mendeley will actually export to BibTex, creating a different .bib file for each collection in my stash! This handy post on the Mendeley blog told me all I needed to know, sort of. Sadly when I looked the pictures were down, so I couldn't follow step by step. The only small problem is that you can't seem to control which collections are exported. But that's tiny. So I've created a Mendeley collection called "bib" (because that was the name of the file in the template, it's as good as any!). Anything I want to cite will get added to that collection, and the exported bib file gets updated. 

What I'm doing is exporting them to a separate folder, then copying my bib.bib file (heh, silly name) into one happy thesis folder. I figure there is slightly less chance of accidentally overwriting them then. 

The next challenge was how to actually reference them in my LaTeX document. This seems to be slightly more convoluted. Fortunately, this post from CUED (blast from my past!) gave me nice simple instructions. Being on a mac makes running BibTeX as easy as opening up a terminal and typing "bibtex {filename}". 

It has added to the number of lines in my included files that I need to comment in and out if I want to compile a chapter separately. I'm up to 6 lines now:


My text goes here


That's still not too painful.

I still have tables and figures to sort out, so I'll update as I sort those (however I do that!). For now though, this approach is still working for me.


Thesis tech

I started trying to write my thesis.

That sounds pretty big and scary right there, especially as I only have half my data at the moment. Still, half my data means I can write a good chunk of thesis, and there's no time like the present to get going.

This means that I sat down and thought about what tech I want to use, given my "writing process" (ha!). I know I like to have a choice of writing locations, and I like using IAWriter on my iPad. I know I don't particularly get on with Scrivener (still not entirely sure why, I think it's the formatting difficulties as much as anything else). I also know I don't want to work on one looooooooong document. I want to be able to send individual chapters to my supervisor, and deal with the responses in a single place (again, I had problems doing this with Scrivener). I also know that my supervisor likes to read and comment on PDFs on her iPad.

What I've finally hit upon is the following: 

  • I'm going old skool, and learning LaTeX. I'm used to HTML and CSS, how hard can this markup be? (Actually, as I've started already "not very" seems to be the answer.)
  • I have an unofficial LaTeX template, and I'm creating separate tex files for each chapter. I add a start and end bit to each document that I can comment out when I want to compile the entire thesis, or leave in to compile each chapter. It's 3 lines.
  • I'm sticking that in a Dropbox folder. Instant backup, plus easy access in IAWriter.
  • And in GoodReader, so when I've compiled on my desktop I can instantly get access to the finished PDF on my iPad to show someone, or email it on if needs be.

This seems to be working for me so far (although granted I'm only 3 pages in). It separates out the writing from the look - I can stick in the markup knowing that it will look ok eventually, and it's semantic enough for me to read it in text form. Equally, reading back and editing seems to be best done from a compiled version. I'm a programmer, I'm used to that kind of set up.

I was sort of thinking I'd wind up testing this on one of the main sections, having seen advice to leave the intro and conclusion to the end. But my supervisor actually suggested starting with the intro, because that kind of sets the framework for what you're doing throughout. You can always go back and edit if your analysis doesn't match. The conclusion should end up mirroring the intro, apparently. We'll see when I get there! I don't imagine for a minute that the rest will be done in order.

Today I used my setup to write up until a meeting with my supervisor, take the latest version in to talk about, get feedback and go back to writing on my desktop almost seamlessly. Fingers crossed it continues to be (technologically speaking) that straightforward throughout!


Fun and Games meta-game

While I'm listening to talks or wandering around at a conference I like to play a "meta-game" of evaluating the slides, presenters, and my experiences, to try and glean pointers for how I can play the conference game better. I do tell people I'm doing this. They've never asked to see what I say, but some do look a little uncomfortable.

In fact, this time I'm going to start out by critiquing my efforts instead.

This conference I went with a really minimalist packing style, at least partially driven by EasyJet's price for hold luggage. I took my normal, lightweight Eddie Bauer backpack as my only piece of luggage. I think it's 18l. I packed an outfit for each day, my wash stuff, and so on. I can say I used everything I brought apart from the notepad, which I only brought as backup for my iPad. I didn't need to, we were given a notepad and pen when we got to the conference, and my iPad was fine. Was there anything I missed? Well, ideally I would had an extra top for the conference dinner. It's nice to dress up (although a little jewellery would have done that), and to be honest it's been really warm and I've got pretty sweaty each day. I know, TMI. There wasn't really anything else I was missing. It worked really well for getting to and through the airport, dealing with the buses, keeping it with me on the last day and so on. I think it was a really good plan actually.

The poster was ok, but I'm going to give myself a "could do better". You can see it below, but sadly it isn't the one that jumps out at you!
5 Sep 2012 12:36
I ended up folding my poster down the middle (due to the hand luggage restrictions), which looked sort of ok, but I think a better idea would be to design a poster that could be split into two separate sheets. Something that deliberately comes apart. They would roll better than the folded poster, and look more deliberate. I also need to make something more visually striking, with less text and more bullet points and pictures. Something that stands out better from a dull background. It's a good challenge. I did have lots of interesting discussions with people who had read my poster - I kind of need to think about what I had to keep explaining. To be fair, I think a lot of people spoke to me rather than read it (too much text I think!).

So, that's my efforts dissected.

In general the presentations suffered from poor lighting - all the rooms were too light, leading to poor contrast on the slides. As this doesn't seem to be that uncommon, I think it's worth trying to find a way to mimic these conditions, just to check the slides are legible. Whether that's reducing the brightness of my monitor or taking my iPad out into the light, I'll try experimenting and checking. I think the biggest problems were pictures that the presenter wanted to talk to that were just too dark, so that's important.

Other than that there was my usual bugbear - too much on the slide. An enormous great table does not become legible when you highlight a cell! Just put the contents on a new slide if you really need to show them, or better yet, put them in the paper and talk around the important bits. There was some nice designs that would have worked well on a one-to-one scale, but as giant slides in front of lots of people the just distracted.

I was interested to see people all presenting on their own machines - if I am to do that and stay minimal I need to do presentations on my iPad. Something to practice on the lab I think! Prezi was an interesting switch - I could actually really see how some of the previous presentations could have used the zooming and moving thing to great effect, but mostly I try to keep my stuff more basic than that. The use of a laser pointer was good if you particularly needed to point something out on the slide (the projection was way above everyone's heads), but I might just fidget with it. Still might be worth investigating. The thing I picked up at my last conference still holds too - dark background with light text works well in too much light.

I'm still pulling together my thoughts overall, but I have a good post-conference buzz going (combined with post-conference knackeredness - I'm thinking lots, but not terribly coherently!). Maybe next time I'll try for a paper...



I have ethics approval. I am now allowed to conduct my experiments...

(Not that they are particularly ethically troublesome. Still, it's progress! All progress is good progress.)


Presenting style

There's a kind of meta game to going to conferences, and that's looking at presentation styles and slide design. Looking back over my notes from multi.player, I have left myself all sorts of comments about the speed of the presentation, the slide design, the slide contents etc. 

More recently I went to dConstruct, which is not an academic conference. It aims to be an inspirational one, with a theme to talk around rather than a specific technology to discuss and learn about. The speakers are professionals who also do a lot of presenting at conferences. I found it fascinating to see the differences in approach.

By and large, the slides were much better at dConstruct (granted at a design-oriented conference you'd hope for good slide design). There was less reliance on them, and less data displayed. In Don Norman's case, he didn't use slides at all and I don't think the presentation suffered for that. In Matthew Sheret's talk, the slides and more particularly how he moved between them (he'd hacked a toy sonic screwdriver) were a key part of the talk, but didn't distract from what he was saying. 

On the other hand, the academics had a much more rigid structure to their talks that in at least a few cases made them much clearer than those at dConstruct: 


  1. Background of topic
  2. Research questions coming from that background
  3. Methodology used to explore research questions
  4. Results and discussion.


Just like a paper, or a poster, or... well, most academic presentation I guess! 

So what have I learnt? From my time with American Express as well as what I've seen at conferences, I try to put as little as possible on my slides. They should not make sense without me there to talk around them. Putting tables of data on the slides doesn't work because a) it's almost impossible to make them legible and b) they distract people from what I'm saying. Use slides to show people things you can't say, so a picture (SINGLE picture, easy to see), not necessarily quotes. Have a structure so you know what your overall story is, although perhaps make it slightly less obvious in the talk than standing there and reading through it. 

Stylistically, dark backgrounds with light text is more robust if your room is too light. Talk at a reasonable pace, not everyone speaks English as a native and there's no need to pound people. Stay on topic - random 'surprise' slides are only really funny once. Test any tech on a machine other than the one you created the presentation on, just to make sure.

I think you have two goals as a conference presenter. You are trying to be interesting and informative (and I mean you, not your slides). If I can crack that, I'll be doing well!