 |
|
| FriendFeed is the easiest way to share online. Learn more » |
Fwd: Science 2.0 Symposium, July 2009 - videos now on vimeo (via http://friendfeed.com/mcdawg...)
Jon Udell: Collaborative curation of public events
Meryn Stol, Jean-Claude Bradley, Kubke and 3 other people liked this
"A decade ago I envisioned what Michael has just so thrillingly outlined for you..."
- Cameron Neylon
Shout out for Greg's project later in the year where students will reprocess Toronto government data
- Cameron Neylon
Point is that _we_ are the contributors of government data - how do ordinary people publish data
- Cameron Neylon
Lots of pinned up posters - the most common system of notification for events, much more comprehensive than what is on the web. Relatively few of the posters have a related page online. No comprehensive index. Why are posters better than the web? Low barriers, but poor searchability.
- Cameron Neylon
Aim is to inculcate a public awareness of what it means to publish data
- Cameron Neylon
He's trying to produce an aggregation system.
- Mark Tovey
So here's an aggregation for Keene, NH
- Mark Tovey
Showing an aggregation of events on services related to specific place.Pulling from eventful upcoming or whatever. System defines area of things to pull in. This is not a datbase, not a destination site, just an aggregation/syndication. Point is any group can publish a feed on their won website.
- Cameron Neylon
Jon likes to make the point that posters on lampposts outperforms the web as a community communication tool
- Steve Easterbrook
- Union of them is more than the sum of the individual parts. So you can publish an information feed directly from your website and you shouldn't have to rely on an information service - in fact I argue that you shouldn't.
- Mark Tovey
There is an underlying problem with understanding this idea of wiring up resources because it doesn't seem to correspond to peoples physical experience of the world
- Cameron Neylon
- I can't get people to understand that it's not a destination - it doesn't contain anything, it has no authority - it just wants to broker connections back to the original sources. There are abstractions involved into a system that goes back to the original system. Abstractions non-intuitive, for a variety of reasons.
- Mark Tovey
- actually, anyone, anywhere, could spin out an instance of this. And start to look for feeds to flow into it. So I've got people who I call curators - and they're essentially people who make a list of feeds.
- Mark Tovey
System for curators to fire up an instance in any geographical location
- Cameron Neylon
- I'm building this on the Microsoft Azure platform - this is going to be very helpful, because right now there are only 17 cities - eventually I want to be able to say, run more instances of the worker role, and then I won't even have to think about it.
- Mark Tovey
not necessarily just geographical but also topical
- Cameron Neylon
- I have a friend who's running a a bio-optics events - she's always trying to figure out what all those events are - instead of centralized service, it's everybody be responsible for your own stuff.
- Mark Tovey
Core values: collaboration, open data, standards, transparency re-use. Respect for authoritative sources. Shared responsibility for curation. Want people to claim authority over their own stuff. Expect the world to link to their stuff.
- Cameron Neylon
- There are a couple of things here that bear more observation - I want everybody to claim authority over that which they properly have authority for. And then I want the world to link to their stuff. And this responsibility for curation - that's another piece. And syndication is a big deal. People are starting to see they can be consumers. I'd like for them to see increasingly that they could be publishers of these things.
- Mark Tovey
The point that some of these ideas that come from computational science need to be core "literacy" components
- Cameron Neylon
- there are people who have noticed ways of thinking and being in the online world who take on computational ways of thinking - these are the kinds of intellectual tools that we really need to be teaching to everyone.
- Mark Tovey
People are aware of an RSS ecosystem to a certain extent at least
- Cameron Neylon
- Greg talked about the book - but I'm looking at another dusty internet standard - I was to say that ICalendar is another RSS (not another NNTP). We have publishers of RSS feeds. And it's fairly well understood that in the eco-system there are things that aggregate feeds. Could be for example the set of blog posts, or twitter - brings all that stuff together -- and what it means to publish into this domain where all this stuff is happening.
- Mark Tovey
Very similar ecosystem in calendars via ICS but much less awareness of that
- Cameron Neylon
And, in fact the tools that are available to people are very different from the usual wave of blogging tools.
- Mark Tovey
And, by the way, all of those same tools, subscribe to these.
- Mark Tovey
- not may people do that, though.
- Mark Tovey
- this is a missing piece of infrastructure that I'm trying to bootstrap, but I'm trying to get the idea out there. Not the only one, but it's important to get out there.
- Mark Tovey
Only recently possible to publish an iCalendar feed to the world
- Cameron Neylon
So, say, I'm a musician, and I want to put out my tour dates. Only fairly recently did it become true that not just in outloook could you export to the internet where it could be subscribed to. And this I call the tyranny of the technologists.
- Mark Tovey
Lots of people publishing event data to the web va RSS but the data is not structured in a way that you can use. Ok for one site but no standard means to pull it all together.
- Cameron Neylon
- Some of use in the tech world have been beating this idea of feeds to the world for the last ten years. But when you look, there's no data in the information feed.
- Mark Tovey
You need to structure data the right way to make it useful
- Cameron Neylon
There's no understanding. The point is to communicate data. This does not do that. What do you do if you want to be the curator in the annex?
- Mark Tovey
- set up a Delicious account - she became rescottaz for prescottarisona.
- Mark Tovey
- 2 - We pretend that there's a ulr called delicious.com/prescotaz/metadata. It's imaginary. You bookmark it. You tag it as metadata. But the system can figure out that it's really ame -- and it turns out that you can put this kind of this - but for whatever reason, I like the idea of expoing peole to this idea in its raw form. So this notion that you could construct a vocabulary - just by willpower and consensus I just think is interesting and important. And there's some more consensus vocabulary going on.
- Mark Tovey
Remarkable use of delicious to bookmark feeds and setup metadata - see Jon's blog for the details but it is very cool stuff - there is a point here in exposing people to the ugliness of the metadata markup
- Cameron Neylon
- so this wouldn't work if everyone in a community had to do this - but this doesn't seem to be the obstacle. But I'm not nervous about this - I'm perfectly open to any other way that people might happen to agree to do it.
- Mark Tovey
- but what I like about delicious for the time being - and the question is - how am I going to get the people who are going to be my group of collaborators - and I actually realized that the perfect solution is friendfeed. And in the theme of the solution, I am actually subscribing to all of the delicious feeds to all of those curators. The bottom line is that when Susan adds that feed
- Mark Tovey
.. to all of her feeds. That's the kind of the thing that makes FaceBook what it is. It's training people to use feeds, file events, in some of sense.
- Mark Tovey
- really powerful because everything else in the environment is a service that has been enabled. It's any post in this blog which is tagged as being about this project. So - where do we get these iCalendar feeds? One of the most fruitful - almost unknown to the creator of the iCalendar feed - you can take that data in handily. But most of the kind of data people think of is not amenable to any kind of syndication. Almost at the dawn of RSS.
- Mark Tovey
- one of the things I may be telling people is that they may be already running this kind of functionality, but they just don't know it.
- Mark Tovey
- but mostly it's much harder than that. For a while I was making use of a service called FuseCal. Mostly it was making use of webpages - and it was making use of what people think of as stuff on the web. And this service called FuseCal did a pretty good job, not perfect, but good. I actually came up with a recipe for MySpace - one of the points about these social networks but people are not incentivised to provide that data.
- Mark Tovey
- if it's too much trouble, but you instead lean on some service - maybe you care enough not to put yourself in that position to allow that to happen.
- Mark Tovey
- this was a little experiment on the web - if you search for those kinds of phrases - you will find an astonishing amount of information on the web. In this case, this was a PDF file - in which the information was in an image. And the date was in a caption of an image. This is not a problem that you're going to be able to throw AI at anytime soon. This is something that's going to require a combination of crowdsourcing ...
- Mark Tovey
Example: chess club, right time, wrong place in the local paper. It was right at eventful.com, though - you could infer that it had moved.
- Michael Nielsen
There's an interesting social challenge here: the chess club doesn't expect to the be the authoritative producer of data. (I.e., you could just pipe it in.)
- Michael Nielsen
- Many of you do not need extra attention - but here's a site which is a receiver of information for Huntington West Virginia - where are people who are looking for information actually finding it. One of the things that I thought would be good would be to display these feeds on television. I just want to talk again about a way of doing ethnographic research on how to deal with this...
more...
- Mark Tovey
Journalism as a form of feed syndication. Wonderful! (Imagine just piping data in directly.)
- Michael Nielsen
- people do not demand - nor is it expected - that they should be the authority for their own feeds. If that was working you would feel you had flow, and the business models available to you.
- Mark Tovey
irrespective of how you decide who is an authoritative source; how do you propagate changes to all the subscribers in a timely way, such that they find out that the event has moved before they show up to an empty room... (that's a social problem not a technical one...)
- Steve Easterbrook
- then I had this painful conversation with my high-school principal. There are just so many things that bother me about this. It isn't just that they're doing this. It's just that they don't undersd what the alternatives should be.
- Mark Tovey
Lack of awareness: what's the difference between pdf information and machine readable?
- Michael Nielsen
- this is a professional person - he's not an unusual guy - I think that people who are in the space of - I think we really don't have a clue about how people who aren't us - see a difference between writing this information into a PDF file, and putting the information into an RSS. This is distressing. And I think it bodes poorly for a lot of initiatives for what's called Gov 2.0 right now. I don't think we're building the basis for that understanding right now.
- Mark Tovey
"don't give me a copy of your information. Give me a link to it instead, so I can respect your authority." (And, it helps with maintaining data integrity.)
- Michael Nielsen
This is one of those non-physical kinds of things - therefore one of those non-intuitive kinds of things.
- Mark Tovey
- Why do people send things, rather than pointers to things? We don't stop to contemplate the extent to which most people don't understand the difference between pass by value and pass by reference.
- Mark Tovey
queries as first class object.
- Michael Nielsen
delicious as the biggest change in his personal information habits.
- Michael Nielsen
A url is actually a query into a bunch of stuff that I had tagged intentionally.
- Mark Tovey
(Aside: I think there's some interesting connections to Roy Fields' work on REST architecture.)
- Michael Nielsen
- I will normally write down a list of things to do, it satisfies something.
- Mark Tovey
- handing a pointer feels different, but more useful.
- Mark Tovey
- by the way you can determine who were the people who were paying attention to that thing.
- Mark Tovey
The delicious link is at http://delicious.com/prescot... .
- Daniel Mietchen
- Lindsay Lohan - publishing topless picture - allows people to link to it - not sending them object 2) and can influence or control the network effects in doing this.
- Mark Tovey
Hilarious but instructive example: Lindsay Lohans' intelligent use of social media to publish topless pictures using a link, not a jpg. (One of many morals might be: if you want people to learn computational thinking, help them use it in a context that matters to them.)
- Michael Nielsen
- This is a guy who just showed up - he popped up at 7:30. And, indeed he had been on-tour - a 20 CD tour - and in doing so, and in combination with the fact that I had been listening to that virtual feed, and those two things came together, and he showed up on our virtual calendar, almost, I'm sure, without him ever realizing it, but only in Keene.
- Mark Tovey
Wants to partner with newspapers, libraries, and schools.
- Michael Nielsen
- wouldn't have worked unless somebody was subscribing. So: this is what I'm trying to do. I'm trying to talk about a few different entities. I'm trying to work with libraries, and I"m trying to work with schools on maybe a few different things. But most of all I'm trying to get these ideas out there in a broad way. It's about these ideas, practices, principles - I don't see nearly this wide an effect in the wider world. I think it could be a big deal.
- Mark Tovey
- "we want and need to be in the cloud, and yet we fear the cloud for obvious and good reasons".
- Mark Tovey
Good question about what happens when services go south. Its a longer term problem we need to figure out.
- Cameron Neylon
(Thinking out loud: federated multi-service backup is an interesting business opportunity.)
- Michael Nielsen
- I want in the cloud a place that is ultra-reliable and ultra-secure. Which means that, if I choose to pay for it, I should be able to to get multi-generational storage.
- Mark Tovey
- Storage is a separate concern from social interaction - Actually I think I should have the choice of storing my photos - I think there should be a choice of interoperable services. I don't want flickr to do my archive, I want them to provide the corollary services of interaction, participation. Separating that idea may be one which will be difficult or uncomfortable for many people.
- Mark Tovey
Q: could you say a few words about the prospect of people putting things out as tags - and what tools might be out there to do that - and what prospects might be out there for getting people in the habit. A: don't think it's a tools issue, it's an attentional issue. Right now there are hubs that people care to appear in. People desparately want to be in those places. Those places need...
more...
- Mark Tovey
"As geeks... we want this to be about tools, but really it's about social ... it's about this other stuff.
- Mark Tovey
Science 2.0 Workshop - 150 registered 102 showed and 88 have stayed to the end. Excellent wrap up by @judell on elmcity
Michael Nielsen liked this
Michael Nielsen: Doing Science in the Open: How Oline Tools are Changing Scientific Discovery
Kubke, Daniel Mietchen and Jean-Claude Bradley liked this
- Terry Tao's blog -"I'm cherry-picking examples, but I could cherry pick 50 examples like this"
- Mark Tovey
"believed to provide complete description of fluids" - very limited success, even after all these years.
- Mark Tovey
"one of the Clay millenium prize problems (1M) is to understand these problems"
- Mark Tovey
"surveys and critiques all the known approaches" to the problem.
- Mark Tovey
"this is not your typical boingboing post". "and he's sharing these thoughts" - quality of this calibre attracts quality.
- Mark Tovey
- that's just one blog - Tim Gowers - more and more examples -- 4 Fields medalists.
- Mark Tovey
- things that couldn't be published in a conventional way.
- Mark Tovey
- you could think of it as a way of scaling up scientific conversation: "it can become widely distributed in space and time."
- Mark Tovey
- also way of making scientific conversation in an important way, searchable.
- Mark Tovey
see also http://friendfeed.com/toronto...
- Cameron Neylon
- "makes it very easy for future mathematicians to benefit" - the ranking in Google should reflect the significance of a contribution. If search engines are doing their job right, should be an accurate estimation.
- Mark Tovey
If you're the kind of person who's just getting into the problem, that may be a very appropriate posts.
- Mark Tovey
- We tend to think of blogging as a very fixed medium -- all of these blogs - it's possible, through imagination to expand the range of the blog medium. There are much more interesting things now going on which are still going on.
- Mark Tovey
- I want to mention one of these things as we speak. The very first iteration happened just a few months ago. If it's near the limits of his ability, it's near the limits of anyone's ability. And they succeeded. Hard mathematical problem that was solved.
- Mark Tovey
- Terry Tao - just last night a polymath blog was set up - if you go, it's polymath projects.org - although it's only been up for 24 hours or so, there's a lively discussion taking place. I'd like to mention just one little observation.
- Mark Tovey
- Michael Adams - added LaTeX to WordPress to entertain himself. (This could not have happened without this tiny little change). Probably quite a number of things that could be added to blogging software - version control, making blog posts citable - there are many others - that's another talk. It becomes manag. They provide ways of filtering this information.
- Mark Tovey
- Garrett Lisi has been doing his stuff online for a number of years. Starting to invest. these very interesting things out there on the web.
- Mark Tovey
- Big picture idea: the critical point is that they're restructuring expert attention - they're focussing in new ways on new problems. Expert opinion is the ultimate scarce resource in science.
- Mark Tovey
- Just want to talk about one radically different idea - we want to talk about markets.
- Mark Tovey
- importance of markets underlined by story about Russian economist visiting economist ?Seabring in London - "we want to adopt markets but who is responsible for the supply of bread to London"? Idea of creating markets, and expert attention.
- Mark Tovey
Innocentive - 400 people downloaded the challenge -- the kind of person who won was the kind of person that you'd like to have win.
- Mark Tovey
"Someone I admire a great deal, which is Danny Hills, has said that problems that seem very hard on a three year time scale are actually trivial at a 50 year time-scale."
- Mark Tovey
Galileo didn't share - sent an anagram to Kepler - it allowed him to claim priority, but didn't have to share his priority. All of his publications were done in books - you were talking about years or decades. So, what happened was, that it went from that to "publish or perish" - has become "the way to get it in"
- Mark Tovey
- It's difficult for us to conceive of an a world like this - but it's the record that you put online over 5 or 10 years which is important to you.
- Mark Tovey
Q: do you see peer-review disapp? A: I don't think we'll see a demise. Look at Tao's blog - what's going on there. He's taking time and attention that he formerly used to publish papers, and he's putting it into this blog. In earlier time, all that effort would have been divided up into more publications. What's happening is that in switching to new tools, that's providing more value...
more...
- Mark Tovey
- in 1660's it took a long time for reputation to be associated with publishing in scientific journals. Henry Oldenberg used to basically prod people to publish. He'd write to Newton and say, basically, Leibnitz has these results... what do you have? Very sadly, one of the reason there's a controversy is that Oldenberg died in the middle of that.
- Mark Tovey
What is going to happen is that it is going to take a long time to be established.
- Mark Tovey
- PloS - has article-level metrics - not sure if it will get adopted by other journals -- other journals are thinking of these kinds of questions - I think we'll see very slow and then the network effects will kick in.
- Mark Tovey
Q: if we got for a not open, or less regulated market. Brian's just asking the question: don't journals have an important filtering fn, and if that's replaced, how is that effectively to be done in the new world. A: Many replies one might make to this. Asked Journalists - Yahoo Topics, or Google? They prefer Google - the algorithmic approach. If you actually look at the information...
more...
- Mark Tovey
A: The fact that you're an author is not that important there. Some people do a great deal of work, some very little at all. It tends to really be a culture of recommendation. There are a few people of those 37 who didn't achieve very much at all. All of the contributions here are out in the open. So there's full accountability.
- Mark Tovey
- Always, when I see letters of recommendation, it looks like everyone did 85% of the work.
- Mark Tovey
Q: So, just following up on that - you gave example of contribution in 37 - there's also a role for for sensemaking - we need to get some language around that sense-making process.
- Mark Tovey
Michael Nielsen - Doing Science Online
Daniel Mietchen and Jean-Claude Bradley liked this
Questions: How many know what a blog is, how many read them, how many of them run a blog (significant proportion)
- Cameron Neylon
Showing Terry Tao's blog "What's new" - and a particularly abstruse and mathematical post
- Cameron Neylon
118 posts in 2008 - often very long and detailed - in fact published as a two volume set
- Cameron Neylon
focus on the post "Why global regularity for Navier-Stokes is hard"
- Cameron Neylon
aspirational goal: I want my blog posts collected as a book...
- Steve Easterbrook
hmmm, might need to lift my game for that, given how bad turning mine into papers is :-)
- Cameron Neylon
post starts with a brief explanation of the problem - in 2d the solution is known, why is it so much harder in 3d, standard answer is turbulence is difficult
- Cameron Neylon
describes a new way of framing the problem, and describes means of attack to those problems
- Cameron Neylon
quality to start - leads to quality in the comments - serious mathematicians batting ideas around - but also moving around to other blogs
- Cameron Neylon
multiple examples of serious mathematicians with blogs - 4 of the 42 living fields medalists
- Cameron Neylon
a mechanism for publishing small striking insights that don't necessarily add up to a formal paper
- Cameron Neylon
a way of scaling up conversation - widely distributed conversation in both space and time
- Cameron Neylon
making scientific conversation searchable online
- Cameron Neylon
see also http://friendfeed.com/toronto...
- Cameron Neylon
shouldn't the best papers be at the top of the google seearch? The blog is. The ranking should refect the significance regardless of the source
- Cameron Neylon
as search results become more personalized the search results will better reflect your interests and expertise
- Cameron Neylon
We tend to think of blogging as a fixed medium but these blogs are expanding the range of what is possible
- Cameron Neylon
Reference to polymath project: Open source collaborative mathematics originally set up by Tim Gowers
- Cameron Neylon
A new site setup last night http://polymathprojects.org already active discussion
- Cameron Neylon
Point as an aside - Wordpress.com supports mathematics
- Cameron Neylon
blogs are just the beginning...open notebook science is another example
- Cameron Neylon
Example: http://deferentialgeometry.org Garret Lisi's brain online
- Cameron Neylon
All enabling filtered access to the new information sources -> restructuring of expert attention
- Cameron Neylon
Expert attention is the ultimate scarce resource in science - money or people is not enough with out the right attention
- Cameron Neylon
Markets are an excellent way to efficiently allocate scarce resources (even given recent events its better than most other approaches)
- Cameron Neylon
How do you create an online market in online attention - the example of http://innocentive.com
- Cameron Neylon
Aren't all these forms of contribution just a distraction. Should you be blogging when you could be writing papers, contributing to wikipedia, instead of writing grants. In short term these are important questions. Often things that look hard on a short time scale are trivial on the longer term.
- Cameron Neylon
Galelio example using an anagram to claim priority without sharing results. Journals took another 150 years to be accepted as the way of publishings scientific literature - close analogy to online media. Hoard the ideas until the paper is complete
- Cameron Neylon
Current tools are just the beginning - almost boring - the interesting stuff is still to come
- Cameron Neylon
Interesting questions on the future of peer review - splitting the notion of "current best thinking" vs "the record" in the future. Question on Reputation Markets - point that in some sense that is what us previous speakers talked about. Will happen slowly, requires network effects.
- Cameron Neylon
David Rich - Using Desktop Languages for Big Problems
- analogy using various drills
- Mark Tovey
hand drill (python) power drill (C++) heavy iron (various acronyms I don't know) - compare to battery operated cordless electric (StarP and other similar tools)
- Cameron Neylon
domain experts are not computer scientists - 4 million people working on desktop systems. Small clusters could be economic and relatively easy to use, but little specific expertise
- Cameron Neylon
"I personally think we're headed back to an era... where compute platforms are going to become more heterogeneous."
- Mark Tovey
(Meta) software vendors -> slows down posting to friendfeed?
- Inward Cologne
Titus Brown - Approaching Open Source Science: Tools, approaches
Michael Nielsen, Jean-Claude Bradley and Ant Beck liked this
Outline: 1) Sucking people into your open source science project, tool choices 2) Automated testing, what it is, why it is
- Cameron Neylon
Should you do Open Source Science: ideological reasons, idealistics reasons, practical reason: people might actually help you
- Cameron Neylon
no, missed the URL and my google foo is failing
- Cameron Neylon
"Close source science is oxymoronic" - slightly tongue in cheek, fist fight to come later...
- Cameron Neylon
Licensing is not usually your choice - more important thing is to get people to actually use your code. If no-one will use it if oyu pay them why put restrictions on it.
- Cameron Neylon
Real question is how do you attract users..."solving social problems with technology" - or at least reducing barriers
- Cameron Neylon
every project needs - a repository, a mailing list, an openly accessible version control system, wiki and issue trackers great if you have time and manpower but you probably don't
- Cameron Neylon
lots of free hosting sites out there than can be used, only major consideration is distributed or not..
- Cameron Neylon
"distributed version control has changed my life. It is completely awesome"
- Cameron Neylon
DVCS is cool because: decouples fro master, mixed blessings (code bombs from comlicated merges cause effective forks), frees you form permission decisions
- Cameron Neylon
process can still be under central control through process rather then forcing it through technology
- Cameron Neylon
"Open Source" vs "Open Development" - issues with loss of control, time pressure on support and management, frank (robust, insulting?) discussion of bugs
- Cameron Neylon
Loss of control as a problem.
- Michael Nielsen
Tradeoff between the problems and the advantages of oepn source developmnet.
- Michael Nielsen
Takeaway: Is it worth using distributed version control? (Crucial question)
- Michael Nielsen
Part II: "There's no such thing as too many corrections"
- Michael Nielsen
Anecdote: "There is no such thing as too many corrections"
- Cameron Neylon
Background in climatology.
- Michael Nielsen
Using reflection from dark part of the moon (which is reflection off earth) to measure earth albedo
- Cameron Neylon
Many thousands of pictures, lots of corrections need to be thought about e.g. atmospheric differences in different parts of the sky
- Cameron Neylon
"How do we know that our code works?"
- Michael Nielsen
Code was applying lots of corrections over and over again - shows that there is little way of actually checking whether the code works properly
- Cameron Neylon
(Aside: Climate modeller Gavin Schmidt has a great article at edge on related topics.)
- Michael Nielsen
For any complex system, it's going to be incredibly difficult to know whether you're seeing code artifacts, or something real.
- Michael Nielsen
what does "works" mean in research anyway - gives what is expected, or what isn't? Or neither.
- Cameron Neylon
Presumably, it means reflecting physical reality :-)
- Michael Nielsen
should be paranoid about code
- Cameron Neylon
Paranoia isn't eno0ugh, of course. Paranoia might makes you less happy, but it doesn't make your code correct. Need good processes :-)
- Michael Nielsen
(Reminded of airline safety engineering, which is a fascinating subject. Problem of coping with lots of small errros in a complex system.)
- Michael Nielsen
Gives Therac-25 as example of deaths due to software errors.
- Michael Nielsen
"My panacea automated testing - another way to make sure your programs work"
- Mark Tovey
"Probably most useful set of test for scientists are regression tests."
- Mark Tovey
Answers question "did my program change"
- Michael Nielsen
"Do every couple of days - if your input is the same, and your output changes, you know why."
- Mark Tovey
This is different from "is my program correct" but at least you know it changed
- Cameron Neylon
Better - do it automatically every few minutes, so you never go far off track.
- Michael Nielsen
"Functional tests are a little bit lower level - you read it in known data and see if it matches known expectations."
- Mark Tovey
should test with tricky data, not easy data
- Cameron Neylon
Use edge cases (hard data) to test.
- Michael Nielsen
"another very nice thing - if you don't use asserts take this away - every programming lang has an assert."
- Mark Tovey
"if this is not true, die horribly."
- Mark Tovey
"trust but verify" - if it's a single statement, he says, verify it too.
- Mark Tovey
"the die unless is a beautiful command" in Perl.
- Mark Tovey
"I've yet to see scientists who write GUI's in the first 90% of their project"
- Mark Tovey
Regression: "just use diffs and shell scripts" doesn't have to be complicated.
- Mark Tovey
"If you run your test suite you will discovered places that are never executed" - specialized tool needed.
- Mark Tovey
"have a bunch of build clients" that run and test your code.
- Mark Tovey
"if you understand the code in scientific research it is by def'n boring." - tongue in cheek
- Mark Tovey
Greg is "more cynical than I am about current programming, and more optimistic than I am about future programming."
- Mark Tovey
For those interested in following along with the Software Carpentry course (http://www.software-carpentry.org) (of which the Science 2.0 event is a part), there's a Friendfeed room set up for live blogging and questions: http://friendfeed.com/softwar...
Michael Nielsen liked this
Mark Tovey liked this
Other ways to read this feed:
