CAA 2021 Cyprus: R package development session proposal

[Joe Roe]

Dear all,

Following on from the discussion in our first meeting about session proposals for CAA 2021, I wonder if any in the group would be interested in organising a session on R package development in archaeology?

The idea would be to build on from previous CAA sessions on R with one specifically about software development, which is in many ways a different genre to using R for analysis. My thoughts were along the lines of a standard session which would include:

1. Talks on new R packages or significant updates to existing ones, with a focus on technical aspects
2. Talks on general concepts in R package development applied to archaeology, e.g. testing and continuous integration, user interface paradigms, base R vs. the tidyverse.
3. Generous time for discussion of future priorities for archaeological package development

I think such a session would fit well with the objects of the SIG, and I know many members have been at the forefront of developing R packages for archaeology in recent years.

Modifications welcome, of course. And this is probably an obvious idea – so do let me know if planning for a similar session is already in the works!

Best wishes,

Joe

[Ben Marwick]

This is a wonderful idea, what s great suggestion! I'd love to join such a session. There could also be workshop on basic package development.

--
You received this message because you are subscribed to the Google Groups "Scientific Scripting Languages in Archaeology" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scientific-scripting-languag...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scientific-scripting-languages-in-archaeology/ae35ad789d7041ba9ff44756d6226edf%40hum.ku.dk.

[Sophie Schmidt]

Dear Joe,

great idea, I'm in!

I love the workshop idea as well. In this spirit we might try to arrange talks according to "depth" of technical focus or steps along the line of package building (e.g. if someone wants to talk about creating a CRAN-package, that'd be probably the most advanced talk), thereby creating an introduction into package building for ppl who haven't dared so far.

This might be an R-idea, but maybe ppl developing applications with different SSLA want to do sth similar? Or if there aren't as many as to warrant their own session, maybe we can merge this? Because whether an application is written as an R package or for QGIS doesn't much matter for point 3 I believe. For the other two points, some technical aspects might be similar, others might differ and a comparison might be interesting... or not?

What do you think?

Best wishes!

Sophie

To view this discussion on the web visit https://groups.google.com/d/msgid/scientific-scripting-languages-in-archaeology/CAKk1%2BdHoFi301iZqpk44SOoCtqABG0sWf-Z67idJU%2Bbhpn3gpQ%40mail.gmail.com.

[Martin Hinz]

I also like the idea very much, especially if it is combined with a workshop. First of all, I was afraid that such a session would be too much a mix between practical questions, especially those of beginners, and more advanced topics. But if we split that up, that might be ideal!

Even if there is nothing to stop us from proposing an R-specific session organised by the SIG, it is nevertheless right that we should also keep an eye on the other languages here. Perhaps we could find a volunteer to demonstrate how software development and scientific publishing can practically go hand in hand, and this person would then, of course, choose the language for the workshop, and in the session, we would discuss more cross-cutting implementation-independent issues?

[Christian Hoggard]

This would be a great session (particularly if a workshop was included!). It'd be particularly interesting to see what develops from the session - if any programs could be made to benefit particular groups or parts of archaeology.

[Joe Lewis]

Dear Joe,

This sounds like a great idea - I'm in!

I think the workshop on package development would be useful - particularly the breaking it down into 'depths'. This would definitely allow for different people/use cases to attend.

I agree with the 3rd point not being R-specific. I think the idea proposed by Martin would work well (especially because I've seen the Julia programming language increasing in popularity).

Kind regards,

Joe

To view this discussion on the web visit https://groups.google.com/d/msgid/scientific-scripting-languages-in-archaeology/71027c1b-c832-4263-aac3-b2992d430785n%40googlegroups.com.

[Joe Roe]

Thanks for for the input everyone. It sounds like we're all agreed such a session would be a good idea, so I will go ahead and start drafting a proposal, pending approval from the SIG to put it under its banner.

Regarding the scope, the idea in making it R-specific was not to exclude other languages but to try and maintain some cohesion. A session on "package development" unqualified could include a large proportion of what is usually presented at CAA, but perhaps that is just a question of wording: "package development for scientific scripting languages" maybe? As Sophie says, it probably comes down to whether there is enough activity e.g. Python development to fill a separate session, or whether we should do it all together.

I also very much like the idea of a workshop alongside a session. And I suppose the scope of that should be left to whoever volunteers to organise it! One thing I'm not sure about is whether a practical workshop would work very well if the conference is forced to move online-only, though.

If it's not too late for everyone to digest this, maybe we can talk further later at the monthly meeting.

Best wishes,

Joe

To view this discussion on the web visit https://groups.google.com/d/msgid/scientific-scripting-languages-in-archaeology/CAM0xMQ8c4iEya6vDYW%3DqBNEtpiamJWqjLfeiWjzB3JquuRp1OQ%40mail.gmail.com.

[Martin Hinz]

I think, that wording or similar is just the way to go! With that we specify that we would not like to see closed source programs to be presented, but rather talk about the process to make software pieces that can be incorporated into different lines of analysis and what to consider when taking up the effort to make it as accessible to others. So I would love to see if the abstract could highlight this and involve more abstract concepts and eg. interface design than how successful an individual software solution is to solve an individual problem. Does this make sense?

[Martin Hinz]

And, I think that although in person might be nicer, nothing can stop an online tutorial to take place. (I kind of practise that at the moment with my students for QGIS, and I assume most of you do that too currently).