Round 12 of Software Carpentry instructor training kicked off yesterday with three well-attended meetings — the notes from the Etherpad are below. We will meet again on Thursday, Feb 19, at the same times (10:00, 14:00, and 19:00 Eastern); between now and then, please do the following:
Read the first two chapters of How Learning Works (and if you picked up a copy of Building a Better Teacher, please try to get through that as well).
Write a short post to this blog on GitHub describing yourself (details are below).
Mail email@example.com when your blog post is ready to be reviewed (so that we know you can send to the list, and so that others on the list have your email address in case they want to reach you directly).
Read the Software Carpentry Code of Conduct. As noted in the meeting, we value the participation of every member of the scientific community and want all attendees to have an enjoyable and fulfilling experience. Accordingly, everyone is required to show respect and courtesy throughout our workshops, online, and this training course.
If you weren’t able to make yesterday’s meeting, or have not had mail from me either pairing you up with someone for Git tutoring or telling you to just go ahead and write your bio, please mail me: the Etherpad didn’t record a couple of people’s GitHub usernames, and a few people weren’t able to make the meeting.
Writing a Blog Post
Previous rounds of instructor training used a WordPress blog. As the class grew, that became difficult to manage, so we are using a GitHub blog this time instead. To add a post to it, please submit a pull request that creates a file called:
YYYY-MM-DD is the publication date (e.g.,
YOUR-NAME is your name in lower case (e.g.,
greg-wilson), and the
.md at the end means the content is Markdown rather than HTML. You
can create this post directly using GitHub’s in-the-browser editor, or
by forking the
repository, adding the file to your clone, and sending a pull request.
If you don’t know how to do this, don’t panic: a third of participants in this round have little or no experience with Git, and another third rate their Git and GitHub knowledge as “middling”. We have paired people up for one-to-one tutoring based on their self-reported Git expertise; if you were partnered with someone, you should have received a message introducing you to them, so please arrange a time to chat online. (We recommend that the novice share their desktop with the tutor via something like Google Hangout, so that the tutor can guide the novice through signing in, editing a page, and so on, but we leave the mechanics up to each pair.)
If you weren’t paired with someone, but would like some pointers, please mail Greg and he will pair you up. Please also mail the teaching12 list or Greg if you would like help, have questions, or would like to make suggestions.
Finally, if you have a few minutes and would like to make a contribution to our materials, please:
listen to this browsercast (which explains what a browsercast is), then
listen to this short introduction to Software Carpentry, and then
look at the other slideshows in the slideshow repository, which haven’t been polished or recorded yet, and either send us feedback or submit improvements.
Thanks once again for signing up - I’m looking forward to getting to know you all.
- Running etherpad in a class helps the students!
- Student’s ask questions in the chat area
- taking notes as a group
- Educational psychology (EdPsych): Study of how people learn. Not
enough to determine how to teach: problem is underconstrained:
multiple ways of teaching are consistent.
- Includes everything from Neuropsychology: what the brain does when we are learning - Physical, in brain
- Sociological - definition of the goals of learning.
- Public school teachers are required to take lessons about
teaching: not true with college instructors, who are experts
in their field but not in teaching
- Like imagining that taking many flights makes a passenger a good pilot!
- Teaching is a learnable skill
- Book recommendation: “How learning works” by Susan A. Ambrose et al
- Question: Broadly speaking - as approaches to teach reading
- Phonics - Bottom up approach: assemble words from letters, then make sentences
- Whole language - Top Down approach: Teach the child to recognize
whole words; not used so much today but rather common in the
- Can start reading whole words sooner, street signs, simple stories, &c
- Good motivation early
- EdPsych: Under constrains teaching–many ways to teach that are
consistent with our knowledge of how learning works.
- Educational Psychology is akin to science.
- Need instructional design: akin to engineering. There are many approaches that are supported by the science of Educational Psychology, but you need to test out what works.
- If you’re using something like an Etherpad or a chat for a class, don’t ask “are there questions”, but offer multiple choices of responses.
- Examples of 2 approaches to teaching how to read:
- Most popular teaching method, bottom up approach to reading, is
- The phonics approach is hierarchical -> could imagine in coding the analogy would be syntax to functions to flow statements
- Whole language teaching teaches children how to recognize words as
ideograms first before explaining particular letters, e.g. “cat”,
“stop”. Top down approach.
- Dr. Seuss used whole language approach
- Most popular teaching method, bottom up approach to reading, is phonics
- Does not matter which method you use - enthusiasm of the teacher is more important than anything else - study in 1952 conducted for the British Ministry of Education
- Method used is insignificant - enthusiasm is most relevant
- Excitement from teacher is something that goes across well. Boredom from an instructor will lead to poor results. Students interpret lack of enthusiasm as the subject matter is not important to learn/know.
- Novelty of method seems to be what determines success, because it is what excites the teacher
- Enthusiasm does not scale.
- “Yet another curriculum” does not necessarily motivate teachers
- Needs to be a sustainable approach to the design of how things are taught.
- When it comes to computing, less Instructional design material available for computing
- Less focus on computational teaching in computing science departments, compared to other disciplines (e.g. language, physics, math) = less material that is directly useable for us
- We’re going to ‘borrow’ results (and hope they work)
- Analogy to language acquisition
- Phonics - giving people the component pieces to write their own code,
- Whole Language - starting learners with small, functional programs and using to develop their skills
- With SWC we are lucky enough to have learners who are already motivated and interested in the material, and it is one of our primary responsibilities not to demotivate them
- Teaching is a performance.
- In software carpentry we have teachers that use both methods - you have to figure out which one works best for you (and for the learners)
- gvwilson will buy you an improv class!!
- 1980’s study by Patrica Benner on how nurses become more competent (books available on Amazon)
- Three type of learners: (full model has more (5), but these are useful enough)
- Novice: (no/limited facts)
- don’t know what they don’t know
- No such thing as pure, uninterpreted knowledge
- no mental model/map
- So novices will often try to apply inappropriate (or wrong!) models
- Like using a highway analogy for the internet
- packing gobs of knowledge into the wrong model - counterproductive
- Don’t have right boxes.
- Questions they ask won’t make sense; they are not even wrong
- Need context (map/model) to ask domain appropriate questions
- Will take a mental model from their experience and try to apply it to the new area - but likely to be misapplying that model
- Our goal is emphatically not to overload with information - because this can reduce how much they learn
- We need to give them a useful model and landmarks of the
- doesn’t need to be complete!!!
- don’t reinforce their incorrect model
- They (might) have something that more or less works, but it might not match reality.
- once they have a model we can fill it with details
- Our aim: Trying to get novices to the point where they can ask
useful questions (ex SO), and then recognize useful answers (on
SO) and learn on their own (e.g. through StackOverflow)
- Not turning scientists into Web developers
- Hard to google if you don’t even know the question (+ keywords)
- this is a “bait and switch” approach to teaching - we offer people “we’ll teach you R/git/Python/etc” and use that to smuggle in a few basic concepts
- don’t know what they don’t know
- Competent: (yes facts, but rigid connection between them)
- Has an accurate mental model of the domain
- Can do routine tasks without significant effort.
- That’s the level we want to bring our participants to
- Recognize a useful answer when it comes up
- Expert: (yes facts, and many, many different connections between
- Experts have more connections between the facts that they know, more direct connections
- Experts can go straight from A - G because of extra linkages,
Competent must go A-B-C-D-E-F-G.
- This makes it more difficult for experts to teach, because they’ve short-circuited the model in a way that a novice can’t yet understand, and when they try and teach it, they include steps that don’t make sense
- I think the point about experts having to fabricate a path to answer very valuable. I struggle with this.
- people become experts when they practice deliberately and reflectively
- We know less about ‘expertise’ than about ‘novice’ and ‘competent’ levels
- Harder to study
- People know facts.Think of these as nodes, which can be linked to each other.
- Expert blindness - forget to “unsee” what they previously saw as competent/novice
- What seems common in distinguishing Expert from Competent is you have many more links between their knowledge - so smaller number of ‘hops’ through your knowledge ‘graph’ to get to the answer
- This maybe the cause of Experts being less good at teaching - they shortcut the reasoning that a novice or competent person would go through
- Seem to have intuition to solve a problem and leap to the answer
- How to get to this level?
- previously thought 10k h of training but wrong
- not enough to repeat but reflecting of what is done
- Reflective / deliberate practice: Critical thinking
- To become Expert:
- Get feedback on work (e.g. feedback to your written texts, at school… etc)
- Start giving feedback to others (book reviews, analysis of existing work)
- The magic happens: Analytic machinery is turned inwards, so
you start giving yourself realtime feedback.
- Reflecting on how to do something
- Ability to perform suddenly spikes upward. Learning curve accelerates.
- Stop thinking about tasks in mechanical terms and start improving output directly.
- Code review is good example for this
- getting and giving feedback
- helps to reflects and to develop expertise.
- Code review can be made to scale.
- Like teaching a child to play piano
- at first she perceives it as a mechanical exercise
- eventually she starts to hear her music.
- Our job (SWC) is to get people from novice to competent. Expertise
is not really part of it.
- Teaching is not conveying information, it is a performance art
- SWC does care about expertise in teaching, but the goal of this course is to get everyone at least to competent.
- Benefits of code review:
- other people will see bugs you don’t
- you will learn new things by looking at their code, and they will learn new stuff from your
- you will learn to critique your code as you write it (giving yourself feedback in real time)
- Teachers spend time in other teachers classroom to watch and learn
(Finland, Poland, Japan, Korea)
- Sharing lesson ideas (culturally ingrained)
- In US - teacher has another teacher in class 6 hours a year; in Finland it is 4 hours a week!
- In improving a school system, peer comments works very well - showing people how to improve is a common thread in successful school systems (e.g. Japan, Finland)
- In healthy Open Source communities, there is healthy interaction and feedback: by teaching other people to code better, the overall project does better.
- Code review: thinking critically about other’s code builds skills and habits to think critically about your own
- Software Carpentry’s github repositories host discussions for similar feedback
- There is a Software Carpentry Code of Conduct both online and offline
- We’re have to make people feel welcome, because if they don’t they’re not going to learn.
- encourage people to sign up in groups, so that you know that there are other people around you who know as little as you do
- having a “wingman” makes you feel more secure in a novel learning environment
- SWC workshops where people signed up in groups, there was:
- better gender parity and also
- better representation of our main target group: people with weak computer skills.
- Some start-ups that aim to give feedback on teaching
- Similar approach was used to learn instruments. Very important to have a healthy, supportive, troll-free community.
- Culture of mutual reflection - give each other feedback - pick good ideas - lead to better teaching
- 2-3% female participation on stackoverflow / github due to hostile, harsh environment (overall industry has around 15% female participants)
- Get people to sign up in small groups. Go with lab-mates/friends.
- Means that you know that other people are on the same level as you.
- People are more likely to show up if people around you are those you can trust.
- Software Carpentry had the experience that if people signed up in group the gender ratio was nearly 50/50 and that more people who were complete novices were signing up.
- Charging for events also cuts no-show rate from 25% to 5%
- Studies have shown that after 45 to 90 minutes the ability to transfer from short-term to long-term memory will decrease significantly. Citation possibly: http://www.unige.ch/fapse/logopedie/files/1914/1285/1086/article1-barrouillet.pdf or http://files.eric.ed.gov/fulltext/ED335141.pdf
- Unfortunately, we are spread out, so people doing the workshops need to
fly around, so for time reasons and cost/travel limitations, we stick to
shorter, intensive formats.
- Feel free to do more if it is local.
- Can also go too far the other way, with classes spread over too much time.
- Four half-days seems to be optimal.
- (what might be done in 5 half days, just to make it a round week?)
- We do not use slides (except for diagrams/schematics - NOT for coding) but simply start coding. People then see that the teacher also makes mistakes (so it’s OK for the student to screw up) and how he/she gets out of the problem. Additional benefit is, it slows us down can take in the information better.
- you can be more responsive than with a slide deck
- Good work re. expertise: Patricia Benner https://en.wikipedia.org/wiki/Patricia_Benner (From Novice to Expert: Excellence and Power in Clinical Nursing Practice)
- Note: use the Teaching12 mailing list to share information.
- Greg Wilson (Software Carpentry Foundation): gvwilson
- What’s one interesting/useful thing you learned this morning?
- Amy Boyle (Washington State University): boylea 0.8
- The usefulness of peer-review applied to teaching
- Kristopher Keipert (Iowa State University): keipertk 0.8
- The brain produces a chemical fatigue response when we “zone out”
- Michael Sarahan (Nion Co, Seattle): msarahan git/GitHub 0.6
- how much enthusiasm dictates effectiveness of teaching (more than methodology)
- Phil Rosenfield (University of Padova, Italy – current time zone
PST): philrosenfield git: 0.75
- experts vs competent and def of expert
- Karl Broman (University of Wisconsin-Madison): kbroman 0.9
- value of peer-to-peer review for learning
- Konrad Forstner (University of Wurzburg, Germany): konrad (git 0.75,
- The three different level of expertise and how to get from on to the other
- Laura Graham (University of Nottingham, UK): laurajanegraham 0.5
- Although it’s fairly obvious - I’d not really thought about the value of the teacher’s motivation for the subject (or way it is being taught) as important before.
- Owen Stephens (Owen Stephens Consulting, UK): ostephens: 0.75
- That teacher enthusiasm overrides method in terms of effectiveness teaching children to read
- Filipe Fernandes (Universidade Federal da Bahia, BR): ocefpaf: 0.7
- Peer review teaching
- R. Burke Squires (NIAID, Bethesda, MD): burkesquires - Github: 0.75
- Usefulness of co-teaching; timing of lessons
- Sarah Stevens (University of Wisconsin - Madison): sstevens2: 0.5
- Peer review of teaching is important for improving your teaching
- Victor (Kwangchun) Lee (XWMOOC, Korea): statkclee
- Motivation and Enthusiam
- Live coding and why not slides…
- Zakariyya Mughal (U. of Houston): zmughal: git/github: 1.
- There is more need for code review and figuring out how to do that and scale is still a challenge.
- Stephanie Mark (University of Toronto): sdlmark - GitHub knowledge:
- I learned that I am probably a novice in the area of my grad research, which explains a lot about why I feel a bit lost
- Dirk Eddelbuettel (Debian, R): eddelbuettel - git/github 0.9
- SWC recognises five learner types, simplifies in on three and focuses on passages from 1st to 2nd
- Pandurang Kolekar (University of Pune, India): pandurang-kolekar
- Hugues Fontenelle (Oslo University Hospital, Norway):
huguesfontenelle (git 0.8)
- phonics vs whole language
- Javier Garcia-Algarra (Universidad Politecnica de Madrid):
jgalgarra Github 0.5
- Definition of expert
- Value of feedback in teaching
- Chibuzor Horatio Emeka (Fasyl Technology)
- John Constable (Sanger Institute, UK); kript git 0.6
- github/stack overflow 2% female contributors :-(
- Sue McClatchy (The Jackson Laboratory); smcclatchy git 0 github 0.1
- live coding as teaching method
- Kyler Brown (U. of Chicago): kylerbrown github 1, git .9
- To attract novices to SWC: invite groups of friends/labmates.
- Pawel Pomorski (University of Waterloo - SHARCNET) : ppomorsk git
0.9 github 0.8
- insight: feedback crucial in writing code
- Ariel Bowers (Space Telescope Science Institute): asymone
- Narayanan Raghupathy (The Jackson Lab): narayananr: github=0.4
- importance of teacher’s motivation on teaching method especially for kids
- benefit of code reviewing
- Kathy Chung (University of Toronto) chungkky git=0
- problem with scaling and motivation; importance of feedback
- Alex Wiltschko (Harvard Medical School): alexbw: git=1.0
- The concept of experts having more links, but not more nodes in their knowledge graph.
- Matt Probert (York UK) mijp1: git=0, mercurial = 0.5
- benefit of code review process to the reviewer as well as the reviewee
- Sommer Abdel-Fattah (McMaster University): sommerab git=0.5
- idea to provide foundation for learning, rather than straight material
- Kim Moir (Mozilla): kmoir - 0.8
- teacher enthusiasm makes the difference in teaching, not method
- Martin Bentley (Nelson Mandela Metropolitan University): mtb-za
(@astonsplat) 0.75 (not done much branching stuff, but familiar with
- Expert-competent-novice distinction in a quantifiable way. Some background info on how SWC works.
- Greg Wilson (Software Carpentry Foundation): gvwilson 0.5 on git
- one thing you learned today that was interesting/unexpected/rewarding/etc.
- Remi Daigle (University of Toronto): remi-daigle 0.5
- Ben Weinstein (Stony Brook University): bw4sz 1.0
- The idea that experts have to create a false path to k
- Elizabeth Wickes (University of Illinois Urbana-Champaign: GSLIS):
- the d
- Morgan Taschuk (Ontario Institute for Cancer Research):
- the path to becoming an expert, through self-reflective practice (and reflective practice can be taught through reflection on other people’s work first). Never thought of it that way before.
- Simon Fraser (Sanger Institute): simonfraser1 git=0.8
- The idea of teaching as a performance art
- Sarah Mount (University of Wolverhampton): snim2 - git/github 0.75
- Positive effects of peer-review on teaching practice
- Violet Zhang (University of Pennsylvania):
- Interesting induction, surprised to know experts are not the good teachers usually, because they tend to jump from the start to the end in one connection, and can’t explain why they take the path
- Donna Henderson (University of Oxford): DonnaHenderson 0.5
- there are cultures where teachers routinely sit in on each others lessons
- Alex Wiltschko (Harvard Medical School): alexbw
- Matthew Bourque (Space Telescope Science Institute): bourque 0.6
- I learned the three types of learners (novice, competent, and expert) as well as the differences between them.
- Matthew Collins (University of Florida): mjcollin 0.6
- method of teaching matters less than enthusiasm
- Evgenij Belikov (Heriot-Watt University, Edinburgh): jevbelikov 0.5
- importance of culture of reflection for teaching expert teachers and the need to still be able to think like a novice to facilitate learning
- Catherine Devlin (18F): catherinedevlin: git 0.8
- I need to look at code reviews & pair programming not as a source of insecurity, but “woo hoo, reflective learning, I’m going to become expert!”
- Johan Hjelm (Technical University of Denmark): hjelmj: git/github:
- the importance of enthusiasm and peer review of teaching
- Hsikai (“Kai”) Yang (U. of Washington) : HKYang 0.7
- the challenge is to avoid demotivating
- Laurie Baker (University of Glasgow): LaurieBa git: 0.1
- Novice students will file learning in t
- Arliss Collins (Mozilla Science Lab): arlissc 0.2
- Fran Navarro-Brull (University of Alicante, Spain) : franktoffel -
- When the magic occurs (self-correction)
- Joseph Long (Space Telescope Science Institute, USA): josePhoenix
- That teacher enthusiasm is the key factor in the success of a particular teaching strategy
- Meredith Durbin (Space Telescope Science Institute): meredith-durbin git: 0.5
- Michael Corey (Facebook): mcorey - 0.75
- How SC promoted gender parity in their workshops
- Rachel Glover (Food and Environment Research Agency, UK):
- The difference between novice, competent and expert and how experts can be bad teachers because they jump from A-Z with little explanation
- Sarah (college): linuxchiq 0.1
- That there is a physiological reason that after 45 mins your brain shuts down. :)
- John Girgis (university of Ottawa - EST Time Zone): Noiraud, 0.4;
Biochemistry research assistant with some experience in R looking to
teach it to more people in my facility
- Teacher enthusiasm is best determinant of learning success
- Adam Richie-Halford (University of Washington): richford - git 0.6
- We need to maintain a culture of mutual reflection in order to deliberately practice teaching.
- John Pormann (Duke University Libraries): jbp4444, 0.5
- Insight: why experts
- Peter Fields (University of Basel): peterdfields: 0.5
- Research support for the premise that enthusiasm for a particular method trumps actual method of teaching in effectiveness.
- Jon Borrelli (Stony Brook University) jjborrelli: 0.6
- becoming an expert requires reflective practice
- Ron Drori (Hamaarag, Israel) ronidrori
- Malvika Sharan (Uni-Wuerzburg, Germany): Malvika: 0.5
- Differences in the mental model of a domain among competents and experts
- Teaching is a performing art.-Greg Wilson
- Peter Clapham (Wellcome Trust Sanger Institute, UK): pc7 0.5
- Reflective learning
- Matt Dickenson (Duke University): mcdickenson 0.8
- Instructor enthusia
- Marios Isaakidis (Mozilla Reps): misaakidis
- Tessa Pierce (University of California, San Diego / SIO): bluegenes
- Method is less important than instructor enthusiasm!
- Alistair Walsh (Swinburne University, Melbourne): alistairwalsh
- The enthusiasm of the teacher is more important than the method of teaching
- John Moreau (University of Missouri-Kansas City): JohnRMoreau - 0.4
- Greg Wilson (Software Carpentry Foundation): gvwilson
- one thing you learned in this session that was interesting/surprising/worthwhile
- Auriel Fournier (University of Arkansas/Arkansas Coop):
aurielfournier (github) @rallidaerule (twitter) (0.85 github) (.2
- the importance of feedback, and how difficult it can be for experts to give the right help/feedback because they have too many connections
- Eli Draizen (National Center for Biotechnology Information):
- Teachers do better when watching other teachers
- Dorota Jarecka (University of Warsaw / National Center for
Atmospheric Research): djarecka 0.8 (not expert on git, but use
- (i’ve learned more about idea of live-coding during workshop instead of presentation)
- Will Cornwell (University of New South Wales); wcornwell 0.7
- The benefits of live-coding rather than presenting previously written code
- Darya Vanichkina (University of Queensland, Australia): dvanic 0.8
- The different issues that need to be balanced when organizing a SWC workshop - how the time/date/hour of the workshop affect who is able to come
- Derek Howard (University of Toronto): derekhoward- 0.2
- There are completely different levels of teaching review in different countries (and not just appraisal).
- Erin Mastrantonio (Education Elements, California): maserin 0.1
- The distinction about how experts have more linkages between facts; explains why it has been so hard for me to teach Argentine Tango after dancing it for many years.
- Christina Harlow (University of Tennessee, Knoxville): cmh2166 1
- lots of small changes, feedback better for open learning communities
- Brian Magill (SSAI, Hampton, VA): bmagill1250) 0.3
- Techniques for achieving gender parity. Maybe this can be applied to other areas?
- Vladimir Sudilovsky (NASA/ADS): vsudilov - 1.0
- swc flys people to go teach?
- Jessica Gallinger (University of Saskatchewan): gallingerj 0
- Tenure history – I will double check! I had learned it was in response to McCarthyism
- David Clarke (University of Western Sydney): TheKingOfThePotatoPeople
- Effectiveness of group sign-up
- Richard Tomsett (Okinawa Institute of Science and Technology):
- Best performing educational systems all seem to include peer observation/discussion; other aspects (e.g. number of exams) don’t seem to affect outcomes as much as this
- Horacio Vargas Guzman (Universidad Autonoma de Madrid/ MPIP-Mainz): govarguz
- Belinda Weaver (University of Queensland/QCIF): weaverbel 0
- That we have things to learn from comedians to improve our teaching - will give it a go
- Daiva Nielsen (University of Toronto): daivanielsen 0.5
- Elegant explanation of why experts make bad teachers.
- P.S. Hi Luke! :)
- Luke Johnston (University of Toronto): lwjohnst86
- Good information on the distinction between expert vs competent. Something I will need to keep in mind as I develop in my skills/career.
- April Clyburne-Sherin (The Hospital for Sick Children): aprilcs 0.25
- If you load too much information into novices before they have a proper framework, it can impede learning!
- Daniel Wheeler (National Institute of Standards and Technology):
wd15 – 0.9
- multiple teachers is so important compared with other techniques for teaching – criticism is key
- Tessa Fallon (AMNH): tessafallon
- Jeremy Gray (university of Toronto): jeremycg0.9
- Anne Henriksen (JMU and UVA).4 the characteristics of an expert
- Richard Kip (University of California, Santa Barbara) rkip
- that the method of teaching isn’t as important as enthusiasm
- Arvind Sundaram (Norwegian Sequencing Centre, Oslo, Norway):
- Lot of (new) facts are helping you to drive the point
- Philip Riggs (US Department of Agriculture): philip-riggs github:0.1
- The top-down versus bottom-up teaching approaches, how they are demonstrated in teaching reading, and how they are applied to teaching programming. Also experiencing group notes using MoPad was eye-opening.
- Avijit Bandyopadhyay (unattached): bandyopa (0.9)
- The differences between novices and experts and about the expert becoming that way due to deliberate practice.
- Andy Haefner (Lawrence Berkeley Lab): ahaefner (0.9)
- Experts have more connections than competent person
- Andrew MacDonald (University of British Columbia): aammd (github)
@polesasunder (twitter) (0.6)
- Ed Psych is Science, and Instructional Design is Engineering.
- Heather Gibling (Guelph) hgibling (0.5)
- the difference in how teachers learn from each other in different countries