Eloise's thoughts and fancies - php

Dr. Who and why the BBC should remain special

nospam@example.com (Eloise Pasteur) — Sun, 18 Apr 2010 13:29:32 -0600

The new Dr. Who (Matt Smith) and it's new head honcho (Stephen Moffat) are 3 weeks in to the run of their season. Long enough we've settled everything down and we're starting to get a good solid feel for the series. Unlike some (you know who you are!) I trusted Stephen Moffat and his choice, and my trust has been well rewarded, at least in my opinion. Three shows is a small sample on which to base such a comment but my honest feeling is that in 20 years time, if we're still obsessing about Dr. Who, it's likely to be David who? Oh, the one before the great Matt Smith! One of the names that only the nerdy will remember.

Trying to analyse why is a bit harder. But, to be honest, I think there's an element of Matt Smith reminding me of both of "my doctors" (Jon Pertwee and Tom Baker) - the ones that played the role when I was a child and that I watched avidly back then. He's got that somewhat manic energy and the ability to interact well with children, to highlight his own childish wonder at it all that Tom Baker had. He's got that somewhat kooky dress sense and sense of dangerous physicality (Venusian Karate anyone?) that Jon Pertwee had. OK, the "Indiana Jones" look isn't quite as OTT as the frock coats and the like of the Pertwee era, but because Matt Smith is so young and the look is quite old there's an element of that to it. Matt Smith's mad hair is like a darker version of Pertwee's too, looking at the picture to the left.

That's not to say he doesn't bring himself to the role too. Bluffing Daleks with a jammy dodger - brilliant!

How does this spill over into the wider BBC? Although he's thankfully shut up, or maybe just squeezed out, the Murdoch empire think that the BBC's special status should be removed - so that we have to go to them for a TV service seems to be the bottom line. The trouble with this was highlighted in some of the other stuff surrounding yesterday's Dr. Who.

The BBC has a wide ranging mandate, built into the system, to educate, entertain and inform. Yesterday's Dr. Who was basically entertainment but without spoiling the show for those of you that haven't had the chance to see it yet, it was set in and around WWII in London and Winston Churchill. It's possible there's a small amount of informing going on there too. However, it has also tied in, and advertised, the fact that there is a huge archive of Churchill's letters etc. just made available online, via their website. Could Sky do this? Sure. Would Sky do this? The evidence to date says no. The BBC does this incredibly regularly. Often it's tied in more to shows with a clear factual content but even things like EastEnders has BBC organised support material and information about reaching support organisation when it does some of its more harrowing story lines.

I'm not saying that the BBC is perfect: there are doubtless many ways it could be improved. But the surety of funding it has, combined with that wide reaching mandate, means it often attaches education to entertainment, more background to informative shows and the like. Big thumbs up from me.

Coda - is it really the last word in coding?

nospam@example.com (Eloise Pasteur) — Wed, 08 Jul 2009 12:56:18 -0600

Coda is a piece of Mac software for coding on a server.

It combines an FTP client with a context-sensitive text-editor with auto-complete and some other nifty bits: a visual CSS editor, and a built-in preview function for example, and plugs-in.

What does that mean in English? Well instead of opening Transmit (or Cyberduck or similar) and using that to organise your files and folders across your machine and the server, but editing in Smultron or similar (Text Mate, Text Edit, Preview etc.) you fire up Coda, connect to the site, and start editing away. You can always the file structure in the site you're working on because it's there in the sidebar. How useful is this? Well, it varies, but if you're crossing linking to other pages in your site, it's awesome because it's easy to check the file names directly. It's useful enough that even on days I don't need it, it's a plus. The iMac has plenty of screen estate to use, but the MacBook likes the smaller footprint this gives, and I like not switching between applications unnecessarily.

It also remembers what you were working on, so if you have been working on a micro-site and open up the main site it diverts you to the micro-site AND reopens all the files you were working on. THAT I find wonderful, knowing I can go for lunch or similar and come back, reconnect and pick things up from where I left it. Lovely touch. Finish that site? Close the files, navigate back to the root, and next time you open that site, you're back there.

Not necessarily to everyone's taste, but it uses tabs for your open files so you can easily (well for a smallish site) have all the files open and flick between them. Smultron lets you do this from the side-bar, but I've got used to tabs, and like tabs here too.

Again not to everyone's taste, but there is a visual as well as a text-based CSS editor. I find I write the CSS in the text-based editor, and usually edit it that way too, but the visual editor lets you see all the classes you've created with their colour coding, so if there's an element that's misbehaving the visual editor gives me a wonderfully quick system to check why that might be.

I'm not normally a fan of auto-completing editors, but this one I like. You start typing the code and it narrows down to find the right thing, using a smart selector, (so if you're in html and writing as an instruction mysql_query() it won't pop up the completion) and in html once you complete an open tag (e.g.

Stock-market prices Histogram

nospam@example.com (Eloise Pasteur) — Mon, 11 Feb 2008 16:42:31 -0700

As someone commented yesterday, my blog recently has mostly been movies reviews. There are three reasons for this:

We've suddenly had a glut of good movies to go and see (well movies we wanted to go and see, even if some disappointed).
I've been ill. Not severely most of the time, but enough that I've been working slowly and taking it gently to make sure I get well
I've been involved in a project that has taken a fair amount of time to polish off.

That project is pictured above. Melanie Swan commissioned a "solid histogram in Second Life to show stock market prices" and I (finally) delivered. There will be a fuller write up of the technical details on my main website soon - I'm going to be developing this, with her continued support, into a more generally usable tool for plotting histograms, and once that's ready, the main website as well as the blog will gain some new pages.

There are some nice bits to point out here though. The script in SL pings a web-server every minute. That server scrapes both the google stock market data for current prices, and the yahoo pages for historical prices, does some fancy footwork and returns the values to be plotted. The script takes that return and converts into these bars. In case, like me, you don't follow stock-markets, the top of each bar is the highest trading price each day, the bottom the lowest trading price. The colour indicates if the overall value rose or fell. The final column (on the right in the image) is a little different - that is today's values, and you can see a little red blip on the edge of it. That is actually a "sleeve" which shows the variation from the opening price to the current price on the same basis, giving you some real time data.

In addition, if you want to see different stocks displayed, you can touch the histogram, say the name of the stock and it will display that stock price for 5 minutes before reverting to the default stocks. (This will only work correctly if the yahoo and google ticker symbols are the same.)

The twiddles? Well the sleeve changed what was displayed part way through, and required some reworking. Displaying data for "today" when the markets were closed was interesting. Displaying nothing before the markets opened proved relatively easy. However, at some point after the markets close, yahoo updates its page to include "earlier today" into the historical data. Working that out to have the google pages input suppressed and the yahoo pages historical data shown hit a couple of bugs and has a relatively ugly, but working, hack in there.

I describe this histogram as 2.5D (a mathematician I spoke to described it as 5D). The height of the bars, and their left-right position are signficant, as are their colours, giving you two spatial dimensions and a 0.5 for colour. One obvious, and hopefully soon to be implemented addition will be to add a third dimension to the bars... their depth, to display additional data (open and close values in the stock-market for example).

Following that, the whole lot will be adapted to make it relatively easy to read data from a notecard, or from a scraper that you write (so there will be a "plotter script" and a "data provider script" that can be changed to suit your needs. Melanie has already made the stock-market histogram freely available as open source from her offices in Second Life. As these additional tools come along, they will also be open sourced and available from Melanie.

Dear diary...

nospam@example.com (Eloise Pasteur) — Sun, 06 May 2007 14:04:36 -0600

It's been another fairly insane week with one thing or another.

Beth is running an open-access American Literature course in SL. I will be attending as an "interested other" rather than studying for credit. The class will mix SL-heads like me (some of whom may be signing up for college credits for it), and RL students from her college, some of who will doubtless be taking their first steps in SL with their first class. This has lead to two things:

Beth and I rebuilt her AmLit classroom. I promise I didn't suggest this (OK, I suggested a garden or forest theme), but Beth was unhappy with the classrooms approach and wanted something open and untraditional.
I started a commonplace, or commonplace book for the course. Given I'm not doing it for credit I have a sneaky suspicion I'll pass on the essay writing - I can hear Beth sighing in relief too, one less assignment to mark each week - but I've blogged about a commonplace book class already and I want to see how it works in practise.

Having rebuilt the classroom into a forest, Beth ran a class there yesterday (to educators rather than students) to see how it went. We had an interesting mix of people show up: One of Beth's former RL students. A Chinese teacher very new to SL, and although with good English not fully fluent. Me, fairly conversant with SL I'd hope, but not conversant with studying literature, nor particularly with American Literature (although there were parts I recognised), another SL-head of US origin so more used to the cannon. She's blogged about it and we've had some inworld discussion about it, so I won't record that here, save to say when I described her as a "grit in the oyster" style teacher, she liked the allusion!

I need to blog about citing and copyright and things. I wrote about it on SLI, but think it could do with an entry here too, where it will have a slightly different tone. I'll add that to iGTD now.

I've also started some pages with learning skills resources. So far there's a General dyslexia page and a writing sentences page. I plan to add some numeracy resources, writing frames, referencing references, "what does this question word mean?", revision tips and notetaking tips pages. I'll happily add others too, if there's a need that I spot, or that you'd like to see addressed. Currently the pages need some work to have an index and probably a column setting for easier navigation. Currently I'm keeping ownership of the pages, but I will consider duplicating them on a wiki for more general access and commenting too. If you'd like to suggest changes or improvements, please comment here, or email me.

Fingers crossed I polished off a land management system for Sudane. It was working smoothly until I realised I'd missed one part from the specs. Changing everything for that has proved... messy. Lots of knock-on effects. Having, in the previous paragraph, commented about exam technique, at least peripherally, that old saw of "read the question properly" is rather strongly in mind. It applies quite well here too... D'Oh.

Finally the photos of the big fight in the Monday night RPG I blogged about have come through. I need to upload them and so on, and will add them to that page, and to a new entry for those that don't want to plough back. I'd do it now, but I have a class in 10 minutes and I need a coffee and some loin-girding first.

More changes...

nospam@example.com (Eloise Pasteur) — Thu, 01 Mar 2007 12:48:57 -0700

Well, the new data structures work, after a bit of tweaking.

php/swf graphs do double Y axis plots, although it takes a bit of tweaking to get it right, you have to draw the second set of labels by hand (well position them within the graph by hand anyway). That said, they plot nicely and you can see them here if you follow the drop down to the last month's data.

The in-world scraper proved to be horribly unreliable. I've switched to a cron-job and cURL-ing the data in. cURL still looks kind of scary to be honest, but I found a nice tutorial, copied it, hacked it and it works just fine. I'm thinking I could add a sidebar for friends online to this blog using the same system, but I'm not sure I'm quite that brave yet. I might just build a scraper for it though, to see how it's done.

I've added links to Tateru's graphs from here, and from SLI to here too. Must remember to check the economy pages for L$ in February sold on LindeX and comment on that in an economy report coming soon.

Data storage

nospam@example.com (Eloise Pasteur) — Tue, 27 Feb 2007 23:24:53 -0700

Well despite feeling like I'm coming down with a cold I managed to hack the storage code to store 6-hourly and daily summary data. I then managed to screw up the mysql_query syntax (I'm blaming the prodrome for the cold) so I lost a few hours worth of data, but it seems to be fixed now. I won't know about the daily data until tomorrow of course, well midnight SL time.

Next step - hack the index page. That's a bit easier because it's not affecting the data so I can just duplicate it and play it through. I'm going to offer a last week, last month and all time graph as well as the current last day graphs, probably with a third graph on the list for transactions through history too.

I checked with the SLI people, and since it's linking out to a data site I'm OK to tell folks about it occasionally. It is a service for them after all.

I need to decide what I'm going to do about back checks. I think on Thursday I'm going to advertise the graphs will be available, and add to that the information that, each month, from April 1st onwards I'll be looking back at the average transactions in the month (LindeX and in-world) so we can see how the economy is growing.

I'm also hoping to ask Lawrence Linden for some extra stats so we can track in-world transactions better (hourly as well as daily data) and track land sales too. Whether or not that will be successful I don't know, but it can't hurt.

I quite like the size of the two graphs I've got... I'm not quite sure what to do. The options I see:
I shrink the graphs by 2/3rds vertically I can fit the third graph in at the bottom and keep the page layout basically as it. A column of graphs, a column of explanations, and a column of links. I'll have to remember to add SLI to the top of the list.
I can keep the graphs and page down for the transactions history (which is less interesting to me than the others.
I can tweak the page to add a "choose which graphs to display" option
I see if I can run a double Y axis plot, it's certainly possible the tool will let me, but I'm too tired to work out how just now.

Overall I like option 4 the most I think. Plotting transactions and exchange rates on the same graph makes a certain amount of sense to me.

How things change!

nospam@example.com (Eloise Pasteur) — Tue, 20 Feb 2007 12:18:58 -0700

Panachart has some nice features, but no way to display legends!

It turns out php/swf charts from maani.us do, and although they take a bit longer to draw they're pretty easy to set up and modify. OK, so I'm a purist, but I like my charts to have proper legends.

They also have the benefit that you can set minimum and maximum axis values pretty easily, so I can actually plot something that shows the exchange rate fluctuations meaningfully - on panachart they're sort of there, but I couldn't get them to work nicely. So, I've switched. Check out my front page for the graphs. Sometime soon there will be different graphs to come, showing the week etc. (although I don't have a week's worth of data yet) and sometime after that a month, and an all time data pretty soon.

Graphing tools in php

nospam@example.com (Eloise Pasteur) — Mon, 19 Feb 2007 09:06:44 -0700

I ended up being lazy and doing a search for graphing packages. I found a load of them, and sorry, don't have all the URLs any more.

I remember jpgraph, panachart, php.swf charts as the three I downloaded and tried.

JPGraph is very highly powered, but the help was confusing - MS standard in fact!
php.swf charts looked interesting, but the help was limited.
Panachart was the winner. It doesn't do all the charts you might like (no pie charts for example) but it's simple to use: just chuck the various data and a few other instructions into arrays, slip the arrays into an object and off you go. Pulling the numbers into php from the database will be harder than this, and it does the graphs I want quickly and easily.

Advances

nospam@example.com (Eloise Pasteur) — Sun, 18 Feb 2007 09:15:30 -0700

So, I finally decided what I would do about the database. A little bit of lsl and a little bit of php later I have a system that stores the statistics that I'm quoting for my pieces on SLI.

I need to let it run for a bit longer to build up the history of statistics, then I will look at running some graphs and things from the site so you can go back and look at the data historically if needed or wanted. Drawing graphs will be a new challenge too: I've done graphics in php before, but never thought about how to code for a graph. I have a sneaky suspicion it won't be as easy as I'd like it to be. I also suspect that somewhere out there I'll find a graphing system for php already written which might save my bacon. I wonder if PERL does it any better? I've never used PERL before but you never know, and it is sat there on the server already.

Between that and sorting out things in SL for a couple of projects, and wondering where a reply to an email is I didn't achieve that much for me yesterday. I did, however, help a friend a little, nudging her to find a solution to a nasty SL rotation problem. It was her work in the main, but at least I pushed her that first step to find the right way and got a big thank you, I rather think she'd been working on it for a while and got stuck.

I did finish a deeply annoying book. It was annoying because there were two stories in there that were interesting, but the temptation to give up and throw the book across the room at a number of points for some very basic science errors, as well as some detailed ones, was very high. The point at which the author totally cocked up simple genetics and the origins of thalassaemia still makes me angry some days after I read it.

Today's agenda: Try to rest again for a bit. Sort out of couple of bits of lsl. Sort out some graphing stuff for the statistics. Transfer more of the old wp blog over. I'm starting to wonder if the best way to do that is to set up my own WP blog here, then import it, but I'm not sure I can be bothered. Cutting and pasting the articles isn't too onerous at least, much better than rewriting them all. If I have the time and energy a film review too.