Bryan's Blog 2005/02
What is a Web Service
Last week I gave a kick off talk at the National Institute for Environmental e-Science (niees) on "What is a Web Service". This was in a meeting on "Developing applications for real-time environmental data" (DARTED).
In that (short) talk, I spent a few minutes on the SOAP v REST wars, because I think the environmental community could be sold a dummy if they put too much work into mis-applying technologies. In particular, I do believe that while there is a place for SOAP-ful web services in some applications (we are using them in NDG), in many cases it's simply not worth the effort. The trick is going to be deciding at project conception whether one needs the full WS machinery or not.
Meanwhile, the war rumbles on ... but here from soap-is-dead via Ryan Tomayko is a key point for us application developers:
If everything is sent over the wire as a XML document that is described by an XSD then it all boils down to how easy you can work with these documents. That is working with XML api's like DOM and XPath. The enclosing envelope should be irrelevant to the concerns of the average developer; it should be treated like just any other transport protocol.
The same article goes on to say:
All that extra machinery provided to support the SOAP envelope is precisely that, extra machinery and has never been shown to improve interoperability. Therefore, in terms of effort, interoperability via SOAP is not any easier than doing it in REST. In fact, its actually more insidious because a developer is all too easily lulled in the fallacy that an object is the same as the XML document.
Which goes to show that this battle at least is about interoperability. I don't know where the truth is here yet, so I'm happy to keep on reading about it. But it's not all about interoperability, we need to add security to the mix, and if I have the same code base at each end of the wire (or not), using WS-whatever and/or SOAP may be easier than doing it all (over again) RESTfully. As I say, I really don't know how this war will pan out, but I do know that NDG needs X509 (proxy) certificates, and XML-signature, and probably a fair hunk of WSRF (or something like it) ... so if we can have it in a SOAP toolkit, we'll use it!
by Bryan Lawrence : 2005/02/28 : Categories computing ndg (permalink)
Internet Explorer REALLY Sucks
I've been aware that my blog didn't render properly on internet explorer. I thought it was probably because my CSS and/or my XHTML weren't standards compliant. So, I've done some testing. My new wikiBNL code and my style sheed have been tested with
It passes both. Now, all newer pages should render correctly, but perhaps I can understand it if the older (non standards compliant) ones don't. And what was the problem? IE didn't like comments in the CSS file (which stopped it being valid CSS, but didn't stop other browsers from getting it right).
I'm obviously not the only person with major problems with Internet Explorer. I think this open letter to Bill Gates says it all.
by Bryan Lawrence : 2005/02/28 : Categories computing (permalink)
The Temperature is Increasing
In my last polemic on the State of Fear, I said I'd put up some station data from elsewhere. Recalling that Michael Crichton has his characters imply that one can only trust U.S. data, I've plotted some data from the Central England Time Series. I don't think there is any doubt about the quality of this data (although of course it is from a very small region of the globe).
I've taken seasonal means, and fitted some trend lines for the last 100,50 and 20 years. The actual figure is here, rather than included directly, as it's rather large. You can see that temperature increases over all these periods, but the rate of increase is itself increasing. There is also a consistent signal in all seasons.
A word of caution. While such graphs are interesting, we wouldn't treat these as evidence of anthropogenic (human induced) climate change without lots of other information ...
by Bryan Lawrence : 2005/02/27 : Categories climate crichton (permalink)
Matplotlib
In my last post, I produced a figure of temperature data. As an exercise in learning something new, I produced it with python and matplotlib. I have to say I was pretty impressed with matplotlib. Although I only played with the tkinter gui and the postscript and png backends, it looks like there are lots of other interesting output types including SVG. I'm certainly going to play with it some more.
by Bryan Lawrence : 2005/02/27 : Categories python (permalink)
State of Fear, pick my timeseries
Following on from this.
Well, with 170+ references, there is an implication that his facts are good, but his attention to detail is not so good, or perhaps we just need to remember it's fiction. Anyway, the Hadley Centre isn't in East Anglia (perhaps he's confused the Climate Research Unit at the University of East Anglia, with the Hadley Centre for Climate Prediction and Research, based mainly in Exeter).
More importantly, from page 84 onwards, Crichton starts showing some carefully selected graphs to make some points. In particular, he shows
Global Temperature, 1880-2003, using GISS data.
Same again, but with carbon dioxide overlaid
Same again, but only looking at 1940-1970.
The timeseries of US temperatures 1880-2000.
He appears to be using these to attempt to make the following points:
A: carbon dioxide and temperature don't always follow each other.
B: the U.S. temperature data is obviously the best in the world, the U.S. is big enough to represent the rest of the world anyway, and since that doesn't show a significant warming until post 1970, then it isn't real.
Firstly A. Let there be no misunderstanding. All other things being equal, increasing global mean carbon dioxide must increase the global mean temperature. If this were not true, we would live on a frozen planet! We know this (No serious scientist denies this as far as I know). However, all things are not always equal, and as we will see there are issues with, particularly, clouds. Also, let's not forget natural variability. Natural variability occurs on a range of timescales, and is bound to result in periods (even decades) where the carbon dioxide signal and the temperature signal will not align. In the long term though, they just have to! The real questions should be:
C: Is the observed carbon dioxide increase manmade?
D: Is it large enough in magnitude for the atmospheric response to swamp natural variability?
E: Are there other parts of the system that might respond before temperature, and produce feedbacks which might stop a significant global temperature increase?
All the available evidence suggests: C: yes, D: yes, E, No. That's why all the refereed literature agree on this point.
Let's deal with his assertion (B) that the U.S. data is the only data that matters. Leaving aside the breathtaking arrogance of this position (no one else can maintain quality recordings!), remember that the land area of the U.S. is about nine million square km. Roughly six percent of the land area of the globe, or about 2.5 percent of the global surface area. So, actually, his assertion that because of the size, it has relevance for the globe is nonsense. So now, let's look at some pictures he didn't choose to show. Here, from the same site as he got his data from are some figures he doesn't want to show:
 |
Hmmm ...
It looks like that when you use all the data, we learn that the warming problem is not the same at all places and all latitudes, and so the American figures are not the whole story - but even they show a significant increase in temperature in the last three decades. Note that GISS explain in some detail what data they use and how trustworthy it is!
Sometime, if I have the energy, I'll put up some time series from some other locations ...
Appropriate Use of Facts: 0, Misuse: already too big to count.
On page 188, we have the following exchange, I've added some parenthetical clause numbering so I can rebut the conversation:
... No one can say for sure (1) if global warming will result in more clouds, or fewer clouds (2) ... global warming will raise the temperature, so more moisture will evaporate from the ocean, and more moisture means more clouds (3) ... but, higher temperature also means more water vapor in the air, and therefore fewer clouds (4) ... so which is it? Nobody knows (5) ... so how do they make computer models of climate ... As far as cloud cover is concerned they guess (6) ... well they don't call it a guess. They call it an estimate, or parameterisation, or approximation (7). But if you don't understand something, you can't approximate it (8). You're just guessing (9).
Oh dear, oh dear. In my disclaimer, I should have said, I'm a dynamacist who specialises in parameterisation. Now this is something I do know something about. But I haven't time for this now. Expect me to come back and comment all these points too ...
by Bryan Lawrence : 2005/02/26 : Categories crichton (permalink)
XHTML Table Alignment
Leonardo purports to produce xhtml1.1, so I decided to do some work on validating my pages (with the new underlying wiki format). It turns out that table alignment was one of two big problems (and a few minor ones) for me to resolve (the other one is that images can't exist in paragraphs, I have yet to solve that one - it requires real work in my wiki parser).
I used to use <center> but that didn't validate (even though it worked). So, given it took some finding (thanks), I've repeated it here: To center a table in xhtml1.1, you need to put it in a div of some sort, and, use the margin and text-align attributes. Here, for example is the piece of my css that works for the "main" part of the Leonardo layout:
div#main table {
margin: auto;
border : thin solid black;
background-color : #FFFFFF;
text-align: center;
}
I had hoped this would fix the Internet Explorer problems I've had (well, strictly, the Internet Explorer problems others have had!) - but it didn't. I'm sorry IE folks, don't blame me, blame Microsoft ...
(The latest version of my wiki and embedhandler that are closer to xhtml1.1. compliance are at corewiki.tgz)
by Bryan Lawrence : 2005/02/25 : Categories computing (permalink)
M2Crypto Woes
Guido (he of python fame), has a rant about M2Crypto. This is a worry from an NDG point of view. In it he also makes the following comment about SWIG as well:
... I've yet to see an extension module using SWIG that doesn't make me think it was a mistake to use SWIG instead of manually written wrappers. The extra time paid upfront to create hand-crafted wrappers is gained back hundredfold by time saved debugging the SWIG-generated code later.
We've only been exposed to SWIG in a big way once, and it was a nightmare. The code involved worked on one platform, not on another, on same days, and not on others. Repeatable problems they weren't ... but to be fair, in that case it might not have been SWIG's fault, but then, we couldn't tell!
by Bryan Lawrence : 2005/02/25 : Categories python computing (permalink)
Orographic cloud in a GCM: the missing cirrus
At last a chance to talk about real science: Sam Dean, myself, Don Grainger, and Darlene Heuff (Dean et. al.) have had a paper on orographic cirrus accepted by Climate Dynamics. The abstract is:
Observations from the International Satellite Cloud Climatology Project (ISCCP) are used to demonstrate that the 19-level HadAM3 version of the UK Met Office Unified Model does not simulate sufficient high cloud over land. By using low-altitude winds from the European Centre for Medium Range Weather Forecasting (ECMWF) Re-Analysis from 1979-1994 (ERA-15) to predict the areas of maximum likelihood of orographic wave generation, it is shown that much of the deficiency is likely to be due to the lack of a representation of the orographic cirrus generated by sub-grid scale orography. It is probable that this is a problem in most GCMs.
Update (May 3rd, 2005): The complete reference is doi:10.1007/s00382-005-0020-9. A link to a personal copy of a pdf will eventually appear somewhere on this website.
by Bryan Lawrence : 2005/02/24 : Categories climate (permalink)
Broadband Frustration
"Broadband Britain" is a wonderful but empty phrase. I work at home a lot, but it isn't always easy ...
I live a long way from my exchange. Currently I have ISDN. I might be within ADSL range, I didn't use to be, but now, who knows? Multiple phone calls have yet to get BT to say anything sensible. I try every few weeks to order broadband from BT, and each time I get lost in a torturous maise of transfers and inaccurate information. I tried again this morning, and may try again this afternoon. (Why order from BT? Because if I can't get 512 kB/s, I want to stay with ISDN ...)
I currently have satellite broadband provided by Micronet Broadband and Opensky/Eutelsat (the satellite is ebird at 33E) and I use BT Midband ISDN for uplink. Because I have a home network, I use an F10 router.
The F10 provides a VPN using ISDN uplink and satellite downlink via osda.eutelsat.net, and all my local IP traffic going externally should go via that VPN. Should I say. The VPN is as unstable as hell, and it appears that eutelsat have a small bank of VPN circuits. I get good service some of the time, but a lot of the time, far too much of the time, I see this:
 |
Far, far too often, this follows:
 |
Note added later in the day, and for once, I caught this message, which must have been to do with the satellite, because it occurred in the middle of some file transfers:
 |
The bottom line is that I have to support fail over to ISDN, but guess what? The F10 doesn't support this without manually changing the configuration! The F10 software is all round poor, for example, you'd expect it to have an option to automatically establish the connection for outband packets (my other ISDN modem can do this). It should be able to support 128 kbit ISDN when the satellite is unavailable (or more usually, when the VPN is unavailable). It should also support optional 128 kbit uplink. However, it does neither, so I have to support two routers, and much mucking around.
When satellite broaband is good, it's very good, but when it's bad, it's useless. And it's bad far far too much of the time.
by Bryan Lawrence : 2005/02/24 : Categories broadband (permalink)
State of Fear, the beginning
Well, I threatened to post my thoughts about Michael Crichton's State of Fear. I finally got to start the book last night, and it's immediately apparent it's going to annoy me as I read it. This book has 170+ references in a bibliography, so he's trying to assert some sort of psuedo-scientific credibility. I have a choice now, just read the book, and give an overall opinion, or try and do a piece by piece critique. The former has been done, so I'll attempt the latter, although I doubt that I will have the time or patience to be very complete ...
OK: a disclaimer. My speciality is atmospheric dynamics, although i am becoming more and more informed on clouds. I suspect, with a high degree of probability, that I am better informed across the issues than Michael Crichton. So I'll bring my own misconceptions to bare. Please email me if you disagree (on scientific grounds) with anything I opine on this topic.
What baggage do I bring to this? As a doctoral student I was quite anti-climate-change. Where is the evidence I used to say ... well fifteen years on, there's plenty. So, yes, I start from being a believer. But I'm a physicist. I'm evidence driven, and yes I am UK government funded, but I run a data centre, there is no impetus for me to be on any side of the debate apart from the evidence.
So far I've made it to page 58, so this is going to be a piecemeal critique, that way I may actually do it ...(The version I am reading is the "Special Overseas Edition", so apologies if the pagenumbering differs from yours).
The first time my sensibilities were offended occurred when he referred to Chylek, et.al., 2004 (Global Warming and the Greenland Ice Sheet). I haven't read the paper, but I have read the abstract (we don't get Climate Change, the topic of financial limits to academic journal access is something for another day). The abstract makes the point that there is a high natural variability in the regional climate, and links the results to a couple of well known oscillations, but reports
Since 1940, however, the Greenland coastal stations data have undergone predominantly a cooling trend. At the summit of the Greenland ice sheet the summer average temperature has decreased at the rate of 2.2 ?C per decade since the beginning of the measurements in 1987. This suggests that the Greenland ice sheet and coastal regions are not following the current global warming trend
Crichton used this in a context where he clearly wants to imply that there is some evidence of a conspiracy by scientists to hide evidence of global cooling that contradicts received wisdom.
Well, firstly, the paper clearly discussed the regional context. Secondly, I'm not glaciologist, but glacial advance and retreat are governed by the balance between accumulation (snow and ice etc falling at the top) and ablation (melting, calving etc). If there is general warming, we expect more precipitation. In some places this will lead to more accumulation than is balanced by ablation due to higher temperatures. In such regions, glaciers will advance (and the regions near the glaciers will cool). I've no idea about the details of this paper, but the way Crichton has used it is duplicitous. What this means is that glacial advance (or retreat) in any region of itself is not evidence of general warming or cooling locally, let alone globally.
From now on, I intend to mark this book by a little score card. At this point we have
Appropriate Use of Facts 0, MisUse 1
(NB: The link above to Chylek et al to the reference uses a doi published at this link. It should resolve properly to the abstract directly, but today it goes somewhere dead at www.springerlink.com where it returns "Bad Request". Shame on you springerlink! Meanwhile, you should be able to get it at the first link in this note.)
by Bryan Lawrence : 2005/02/19 : Categories crichton (permalink)
DataDeluge
I have just got a hard copy of the JISC briefing paper on the Data Deluge: Preparing for the explosion in Data.
There are some good thoughts in here, and some key questions:
... e-research data are likely to be annotated automatically and stored in a digital archive or library? What will the role of libraries be in this context?
Indeed. As we all become responsible for our own information environment, my take on this is that the library function will become an advisory function because:
... Could some institutions act as repositories for scientific or technical data on behalf of a number of institutions?
Is likely to be more and more true ... the cost of having local copies of all relevant materials (be they physical or digital) will become prohibitive. However, the cost of not having local expertise in how to find and utilise digital (and physical) objects will be too great for institutions. The library function will remain, it'll just be different. Librarians like to call themselves Information Specialists, and it's going to be true!
By the way, I also liked these useful quantifications of thousand fold increases:
| 1 megabyte | A large novel |
| 1 gigabyte | Information in the human genome |
| 1 terabyte | Annual world literature production |
| 1 petabyte | All US Academic Research Libraries |
| 1 exabyte | 2/3 of annual production of information |
by Bryan Lawrence : 2005/02/14 : Categories curation (permalink)
Testing, Extreme Programming and Demos
Patrick Logan has some really good thoughts on how to do design a demo so that it doesn't break ...
I liked his comparison with extreme programming. I liked this so much, I'm repeating it here, because I figure I'm going to want to find this in ten years time, and who knows how long blogspot.com will last? (But please read his copy not mine).
by Bryan Lawrence : 2005/02/13 : Categories computing (permalink)
Sunday Morning Procrastination
It's Sunday morning, and outside can't make up it's mind whether it is a hailstorm, a gale, rain, or sunshine ... clearly time to do something useful. So I'm procrastinating and catching up on blogs and such.
Patrick Logan's good value this morning. He also pointed me at Sam Ruby's 2004 presentation on encodings and difficulties with URIs. Patrick's quote from Sam that grabbed my attention was:
The accuracy of metadata is inversely proportional to the square of the distance between the data and the metadata.
but the entire presentation is food for thought if you care about standardisation on the web in terms of encoding and why it matters. (Of course, initially I just liked the quote for it's wider connotations).
He (Sam Ruby) concludes with
Comparing characters and uris is surprisingly more difficult and important than you might otherwise imagine (think: security holes).
Having found this source of presentations, I also liked the 2003 presentation on some elements and details of atom, and in particular, this statement on required versus optional elements:
The number of optional features in XML is to be kept to the absolute minimum, ideally zero. As a result of this, any XML document has a high probability of being handled successfully by any XML processor.
... good advice for us in some of our NDG schema, many of which are very complicated, but they don't all need to be!
by Bryan Lawrence : 2005/02/13 : Categories curation xml (permalink)
Upgraded to v166
This web site is now running a very slightly modified rev 166 version of leonardo. I only had to make a few changes:
modify my menu page to use the new calendar insert option
copy my modified wiki04 over top of James ...
add one line to the page template to get an image
copied my version of static.py (which handles a few more mime types and slightly different default file locations) over top of the svn version.
fix a whole heap of bugs in my wiki code. (A revised version is at corewiki.tgz).
At this stage I haven't yet had a chance to test all the functionality of my new latex wiki code, as I have yet to install dvi2bitmap or source-highlight ...
2005/02/12 (permalink)
yet more on licensing
The licensing saga isn't over yet. I so don't want to be thinking about this ... As Steven Vaughan-Nicholls writes in e-week.com, commenting on the plethora of open source licenses:
You shouldn't need to be an IP attorney to be a developer, and there are times that I wonder if that's where we're going.
Anyway, my musings on sundry GPL issues are obviously timely. I should have been aware of the article on a GPL3 in eweek from November last year, but yesterday's eweek article Rewriting the GPL no easy task brought it to my attention. Another useful thing from yesterday:
It (GPL3) is designed to be a copyright license that works all around the globe, so whoever rewrites it had better put on their global copyright harmonizing hat in order to do the task productively.
Wonderful. Meanwhile, for me, the software saga continues, because as well as my curation requirements, we want to release some code into the wild, now. You may recall that in this context my major issue about licensing came down to avoiding liability. Fortunately, better trained minds than mine have been on the problem and crafting a CCLRC license. Hopefully in the longer term we can use a license better understood by the community. Meanwhile, the key parts of it are:
All warranties, conditions, terms, undertakings and obligations on the part of CCLRC, implied by statute, common law, custom, trade usage, course of dealing or in any other way are excluded to the fullest extent permitted by law.
Subject to condition 4, CCLRC will not be liable for:
any loss of profits, loss of revenue, loss or corruption of data, loss of contracts or opportunity, loss of savings or third party claims (in each case whether direct or indirect);
any indirect loss or damage arising out of or in connection with the Software;
any direct loss or damage arising out of, or in connection with, the Software
in each case, whether that loss arises as a result of CCLRCs negligence, or in any other way, even if CCLRC has been advised of the possibility of that loss arising, or if it was within CCLRC's contemplation.
None of these conditions limits or excludes CCLRC's liability for death or personal injury caused by its negligence or for any fraud, or for any sort of liability that, by law, cannot be limited or excluded.
(Note that because of limitations of formatting in this wiki, the clauses are not strictly as will appear in the license, but you get the drift). I wish I understood how this was better than the GPL ...)
In terms of the wider issue, of derived works, and how that concept is embodied in the GPL, we have more work to do :-).
by Bryan Lawrence : 2005/02/03 : Categories badc curation (permalink)
non rigid metadata classification
It is an act of faith in the data management community that we need controlled vocabularies that interact using thesauri and ontologies. It's an act of faith at the moment because we dont think we have any other options ...
This is impacting on me right now:
In the NERC DataGrid development, we are developing metadata schema, and we want them populated with terms that we can use to allow scientists to search on. We know we have to be very careful here though, one persons aerosol content is another persons sulphate (fine), and both may use different units. We propose to address this in general by the use of standard names as well as long names, (e.g. the CF convention and the BODC parameter dictionary). There are many other places where controlling the vocabulary and/or categorising the entries will help.
If we take CF as a specific example, we already have a problem with the fact that the CF convention, and the inherent standard name table are a moving target. Without version control, one doesn't know for sure that the item you point to is the one you want. Even worse, at the moment we have a (perceived) problem that it takes time to get a standard name accepted ... perhaps leading to problems in use of the CF standard names more widely.
However, we have the problem what it is damn hard to get anyone to conform to any standard (who has time to read these things?). Of course, the data centre professionals do ... but we are dependant on the scientists in the first place, and mostly they don't.
i d e a n t has an article entitled Bookmark, Classify and Share: A mini-ethnography of social practices in a distributed classification community. It's about tagging information in the (primarily) blogging community. As he puts it:
... allow users to collaboratively organize a shared set of resources by assigning classifiers, or tags, to each item. The practice is coming to be known as free tagging, open tagging, ethnoclassification, folksonomy, or faceted hierarchy (henceforth referred to in this study as distributed classification), and is associated with popular online services such as furl, del.icio.us, or flickr ... One important feature of systems such as these is that they do not impose a rigid taxonomy. Instead, they allow users to assign whatever classifiers they choose.
There are obvious problems with this approach (can any of us agree on a common system of organising anything?). Jon Udell wrote that:
Conventional wisdom holds that people will never assign metadata tags to content. It just isnt on the path of least resistance, the story goes ...
So, no difference from the science community there ...
He goes on to say:
yet somehow, users ... routinely tag content, and those tags open new dimensions of navigation and search. Its worth pondering how and why this works ... Abandoning taxonomy is the first ingredient of success. These systems just use bags of keywords that draw from and extend a flat namespace. In other words, you tag an item with a list of existing and/or new keywords.
The reason why this works is because
Feedback is immediate. As soon as you assign a tag to an item, you see the cluster of items carrying the same tag. If thats not what you expected, youre given incentive to change the tag or add another. If your items arent confidential and online-only access is sufficient, this can be a great way to manage personal information. But the real power emerges when you expand the scope to include all items, from all users, that match your tag. Again, that view might not be what you expected. In that case, you can adapt to the group norm, keep your tag in a bid to influence the group norm, or both.
By contrast, both James Tauber and David Megginson are suggesting the use of Wikipaedia as a way of providing tagging. In brief, what they suggest is that by referring to a wikipedia article, one defines what one is talking about, or at least defines it in terms of the definition that existed when one made the link. While I can see the simplicity of the idea is attractive, my worries about impermanent links apply here too. Mind you, now I've introduced two concepts:
The concept of categorisation, and
The concept of definition.
While the examples that David and James use for categorisation are probably going to be reliable enough in the long term, it may be that for more controversial topics, the actual categories will drift in time, linking things together in unforeseen ways, and introducing new misunderstandings as the definitions may no longer agree. In my CF example above, I spoke of the importance of version control, and I think this applies to serious application of categorisation ....
Tim Bray has discussed the issue of hierarchies in the use of tags (categories). He introduces us (well me) to the Atom category concept, and I can see the direct cross over to the GML dictionary syntax which we are already building into NDG. The important point is that one uses a URI, either directly, or indirectly in a namespace sort of way, to ensure that the vocabulary origin is understood.
Nearly all the bloggers on this subject speak for simple approaches, perhaps like using wikipedia, or using pings, or del.icio.us or ... The requirement for simplicity applies to scientists, but I think thus far the blogosphere has two further things for us (the data management community) to learn:
the importance of instant feedback
the attractiveness of a reference source that can be manipulated if it doesn't have what you want in it
I began by introducing some areas where categorisation are really important to me (and my community). For us, we have no choice but to have complex semantic markup (we want machine understandable metadata), but we do have a choice about how we get it. To that end we need
A vocabulary server, so you can find definitions (this is under way at the bodc as part of the ndg project,
A way of finding what has linked to particular vocabularies. (We'll be able to do this using our NDG search tools),
A timely (i.e. practically instantaneous) way of users proposing new definitions, and a responsive governance system for deciding on them (yes, we do need that, otherwise we have conflicting definitions, and a number of folk - e.g. see links from here - have correctly identified how important that is).
The third of these is the only thing that we haven't yet tackled decisively, but it's on the menu ...
by Bryan Lawrence : 2005/02/03 : Categories badc curation ndg metadata (permalink)
there is no such thing as a foolish question
Generations of young children have been admonished with variations of a quote typically attributed to Mark Twain:
Better to keep your mouth closed and be thought a fool than to open it and remove all doubt
I say generations, because the same link points out that it probably arises from Proverbs 17:28:
Even a fool, when he holdeth his peace, is counted wise: and he that shutteth his lips is esteemed a man of understanding.
In my experience this is a piece of poor advice that has stood in the way of understanding and progress! I would say that the only foolish question is the one you are too afraid to ask, and the only thing foolish about it is that you are afraid of being held to be foolish! In every educational situation I can think of (and I can think of many), there is almost no time when it makes sense to sit wondering what someone means when they are talking (OK, there is one, if you've been too lazy to read some preparatory material), or having asked a question, be afraid of following up because you didn't understand the answer. How foolish is it to keep your mouth (keyboard) shut and never know the answer to your question?
2005/02/03 (permalink)
Copyright and Blogs
Oh well, while I'm on the legal thing, it turns out that my thoughts on coyright and blogging preceded (coincidently) a rather good article on some of the issues in informationweek.
2005/02/03 (permalink)
Impermanent Links and Icebergs
The wonderful think about the Net is that you can link to things. What's not so wonderful is when the link disappears. Even less wonderful is when the content at the link target changes after you made your link. A couple of weeks ago I linked to NASA article about iceberg/glacier collisions. At the time of writing, the article predicted a collision. As I type this, if you follow the link, it describes the iceberg going aground on a shoal a couple of miles short of the glacier. Who knows what it will describe when you follow the link ...
To be fair to NASA, under "related links" they do provide what I suppose was the original article. However, in general I think this is the wrong way around. If you are going to call something an article, then you should not change it ... except in ways that preserve an information audit trail with a clear concept of time's arrow (in this case the older file is called ice_berg_ram2, compared with the newer ice_berg_ram.html). Far better, to leave the link as it is, and have an "updates link" added to the page ... i.e. to have what would I suppose be a manual trackback ... and have a new article there. The NASA behaviour compares poorly with the blogging concept of permalink. With a defined concept of a permalink, one can have a site like mine: a mixture of permanent things (ok, I admit to fixing gross errors of spelling and grammar), and things that change (e.g. the personal wiki pages and/or external links). It should be obvious to the casual reader which is which and that's fine. What's not fine is to have something which appears permanent ... i.e. an article ... and have it changing ...
Back to the iceberg. It looks like it's gone into reverse, and may be cracking up a bit ... not nearly as exciting a grand crash as seemed likely before, but the images from the NASA site are still a great advertisement for earth observation.
by Bryan Lawrence : 2005/02/01 : Categories environment curation (permalink)