|Steven Grimm cb7f4df624 Link to video of JMS discussing O'Hare's departure||6 anos atrás|
|Production||12 anos atrás|
|background||6 anos atrás|
|chars||6 anos atrás|
|comic||6 anos atrás|
|countries||6 anos atrás|
|credits||6 anos atrás|
|find/CompuServe||8 anos atrás|
|ftp||6 anos atrás|
|gif||6 anos atrás|
|guide||6 anos atrás|
|internal||6 anos atrás|
|making||6 anos atrás|
|maps||6 anos atrás|
|misc||6 anos atrás|
|nav||6 anos atrás|
|novels||6 anos atrás|
|p5||6 anos atrás|
|reference||6 anos atrás|
|resources||6 anos atrás|
|src||6 anos atrás|
|synops||6 anos atrás|
|universe||6 anos atrás|
|.gitignore||6 anos atrás|
|94.whatsnew.html||6 anos atrás|
|95.whatsnew.html||6 anos atrás|
|96.whatsnew.html||6 anos atrás|
|97.whatsnew.html||6 anos atrás|
|98.whatsnew.html||6 anos atrás|
|99.whatsnew.html||6 anos atrás|
|Makefile||8 anos atrás|
|README.md||6 anos atrás|
|all_eptitles.gif||12 anos atrás|
|edl.html||12 anos atrás|
|edldesc.html||12 anos atrás|
|episodes.php||6 anos atrás|
|eplist.html||12 anos atrás|
|footer||6 anos atrás|
|gen||12 anos atrás|
|help.html||6 anos atrás|
|index.html||12 anos atrás|
|jarrell.txt||12 anos atrás|
|lurker.html||6 anos atrás|
|renew.html||6 anos atrás|
|search.html||6 anos atrás|
|toc.html||6 anos atrás|
|volunteers.html||12 anos atrás|
|vqt.html||6 anos atrás|
|welcome.html||6 anos atrás|
|whatsnew.html||6 anos atrás|
There are three main components to the Guide. I’ll refer to them by their local URLs on www.midwinter.com.
The first section, which is the bulk of the site, is a directory tree of static HTML files with a few server-parsed HTML files (for server-side includes) thrown in, plus one short PHP script.
Inside that directory are a bunch of subdirectories, mostly corresponding to the major sections of the site. I’ll get to them individually in a moment, but first a word on how some of the files get generated.
The biggest thing to know about is a Python script called “gen.py”, a version of which lives in each of the directories that contains per-episode HTML files. The script is responsible for assembling a static HTML file out of a header section (generated for each episode; it includes the episode’s title among other things), a body file, and a footer section.
The body files contain the meat of the pages. They have the same filenames as the HTML files, minus the .html extension. In most cases that’s just the 3-digit episode number, e.g. /lurk/guide/056.html contains the body file “056” from the guide subdirectory. When I want to add a new comment to the page for episode 33, for example, I cd into the guide directory, edit the file “033”, then run “gen.py 033”. The result is a new version of 033.html, with the old version moved aside to #033.html as a precaution.
One other thing some of the gen.py versions do is add timestamps to items. The idea is that when I modify a page, I want to highlight the change so people who’ve already read the previous version can quickly scroll to the new stuff. To that end, gen.py looks for a special token “@@@” (which must by followed by whitespace or end-of-line) in the body file. When it finds the token, it replaces it with “@@@nnn” where nnn is the current timestamp in UNIX time_t format. This change is made to the body file itself. Then, when gen.py generates the HTML file from a body file, it inserts the date in bold wherever it sees a recent timestamp. Once a day a script regenerates all the HTML files so that the date markers go away after a while (keeping them in makes the pages look ugly and makes it hard to quickly locate new items).
When you see a file “skel”, it’s a prototype body file. I copy that to the appropriate filename when a new episode/novel/etc. comes out.
In many of the subdirectories you’ll see a “genall.csh” script. This is just a shortcut that runs “gen.py” for each episode. I only use this when I make a change to the structure of the page headers/footers and want to apply it everywhere.
Also, in the top-level directory there’s a “gen” script. This runs gen.py in each of the subdirectories. I use that when, for example, I add a picture for an episode and want it to be displayed on all the pages for that episode.
“The story so far” for each episode of seasons 1-4. Season 5 still needs to be written. The user-accessible files here are all .shtml files. You’ll notice there aren’t any body files here. Instead, background/gen.py uses a few other source files. There are several different introductory summaries, which are in the sum-* files. For example, sum-66 summarizes the story up to and including episode 66, and is included as the first part of the backgrounds for episodes 67 and later. After the appropriate sum-* file is included, gen.py looks in story.html. That file contains per-episode paragraphs that are included in the “More recently…” section of the background pages, under the introductions. Only the paragraphs for episodes between the summary and the current episode are included, e.g. for episode 75, gen.py includes sum-66 and the paragraphs for episodes 67-74.
Bios of some of the characters. These are linked to from the “Universe” section. Simple flat HTML files.
Guide pages for the comic series. Body files and gen.py.
The only HTML file directly under here is index.html, which is a list of all the countries for which I have schedule information. Each country has a subdirectory under here (the directory name is generally the country’s 2-letter country code). One pseudo-country of note is “master”, which contains the master episode list with the episodes in the correct order (the original US airing order was slightly wrong storywise).
Inside each country’s directory is an eplist.html file with the schedule for that country, plus symbolic links to the various episode-guide subdirectories. By using relative paths in my URLs, I let the browser remember which country’s schedule the user was looking at without resorting to cookies (which didn’t yet exist as a concept when this site first started!) For example, /lurk/countries/se/eplist.html has a link to “guide/056.html”, which is really the same file as /lurk/guide/056.html since /lurk/countries/se/guide is a symlink to /lurk/guide. But since the browser doesn’t know that, 056.html’s link back to the episode page is “../eplist.html” which goes back to the Swedish schedule.
index.html is a link to eplist.html in the country directories. There is a “mkcountry.sh” script to set up a new country directory. There are also skeleton episode lists for all 5 seasons (skel, skel2, skel3, etc.)
When I edit a country’s schedule I use the “print-weeks.py” script, which is described later, so I don’t have to hand-type dates.
A short PHP script that does a simple lookup of the user’s domain name to figure out which country’s episode list to show. Perhaps less relevant now that the show is pretty much off the air, but when different seasons were showing in different countries simultaneously this was handy for users.
A symlink to the US episode list, so that any links to “../eplist.html” in the the non-country-based guide directories will go somewhere useful. (See the discussion of the “countries” directory above.)
The standard page footer used throughout the site. Not used programmatically, but I load this into my text editor when I create a new page.
Script that runs gen.py for an episode in each of the episode-related subdirectories.
Pretty self-explanatory. All the non-navigation-related images live here. (Some of them are JPEGs despite the directory name.)
The episode guide pages.
The site FAQ. It’s referred to by the name “help” instead to avoid confusion with the Usenet B5 FAQ.
Symlink to lurker.html.
Data files used by gen.py and CGIs:
List of episode guide and synopsis pages, ordered by modification time. Generated by the “lastupdate.py” utility.
Home page without the black background, since some people find custom backgrounds annoying. This is automatically generated from lurker.html via a cron job that runs “make” in the top-level directory.
The home page.
Once an hour this Makefile is used to rebuild lurker-nobg.html, and once a day it’s used to make sure I haven’t forgotten to run gen.py on something. It also updates the b5tvlist.txt file in the FTP archive.
The “Making of B5” section of the site. Flat HTML files.
Imagemaps for navbar.
Miscellaneous documents that don’t fit elsewhere, e.g. a blurb about why Claudia Christian left the show.
Navigational images. A secondary function is that its index.html is the “preload your image cache here” page, which speeds up the site for folks by letting them suffer all the pain of nav image downloading in one shot.
Novel guide pages. This section is incomplete -- there aren’t pages yet for the latest round of novels.
Results of the Poll 5 viewer survey for each episode.
Button and nav images in their original lossless-compressed forms.
The “Reference” section of the site. Mostly flat HTML files, but a couple of the pages have server-side includes that bring in text files from the FTP area.
Not referenced any more, but perhaps of historical amusement.
The “Other Resources” section of the site. Flat HTML files.
Contains some random support code. More on this later.
The “Universe and Characters” section of the site. Flat HTML files. Season 5 still needs to be written up.
An exercise in naivete; I ended up doing all this myself!
JMS message about B5 being ranked #1 by Viewers for Quality Television.
Introduction to the site after the switch to the current look and feel.
The What’s New page. When this gets to be big, I put all the old items in archive what’s-new pages and link to them at the bottom of this page.
midwinter.com used to host an FTP archive with text files and images; these are now all accessible via the Lurker’s Guide and are stored in the “ftp” subdirectory.
See the FTP index file (available in both text and HTML) for descriptions of the various files in the FTP archive. The Lurker’s Guide points to various files here -- including some server-side include references.
There are a few utilities whose source you’ll find in /lurk/src unless otherwise noted. Some have been discussed above, but here’s a list:
This is run once a day from a cron job. It regenerates all the guide pages. If a regenerated page doesn’t differ from the old version, the old version is moved back into place to preserve its modification date. The purpose here is to expire the timestamps on new items in the guide pages (see the discussion of the gen.py script above).
Run once an hour from cron; this updates /lurk/lastmod.html and is pretty self-explanatory.