Grapheel’s IRIS online description service has now completed its first testing period, meaning that the service is now closed for the time being. The testing has given us invaluable feedback and we would like to sincerely thank each and every one of the kind people who helped volunteer during it. From it, we have discovered lots of ways in which we could hopefully improve and are excited to start working on the next generation of IRIS, which will come online sometime later this year (after summer).
One of the main things that flagged up as something that needs to be addressed is quality insurance of people’s descriptions. We would hate for a user to upload their image and receive a sub-par description or worse. That is why our next big target is tackling the question of how to improve the volunteers’ descriptions of images — and make sure they stay improved. We also need to ensure that the users’ uploads meet a certain standard, e.g. the image is successfully uploaded, legible, and not too complicated; however, that is not as much of a priority at the moment since problems on that front is currently a less regular occurrence.
In all fairness, it would be unrealistic to expect every single user to be completely sympathetic to what needs to go into the description right off the bat. Humans take a lot of things for granted, and sight is no exception. When something is very obvious to you, it is easy to forget what is and isn’t obvious to someone else. So, many users will end up glossing over important details in their descriptions — whilst the reason behind it may be forgivable, it is certainly not forgivable once their description is uploaded. We want to help IRIS users by delivering the best possible descriptions for the images they provide, and that means being strict with what makes a good description and a bad description.
For this, we have created a tutorial for volunteers to complete upon registration. The tutorial gives some examples of different kinds of scientific images, and explains what should and shouldn’t be mentioned. It does not take long, and it is a one-time thing. This is where you, dear reader, come in: to see if it is helpful, we need people to complete the tutorial.
We are enormously grateful for anyone who could take the time to do this, as collating data to see if the tutorial helps is key to refining it. As mentioned, it does not take long, and it is open to people of all kinds of STEM backgrounds — including no background. We would also encourage you to share this with your friends, as the more people that complete it, the more accurate we can be in our analysis.
Thank you very much in advance, and stay tuned for further Grapheel developments!