Hi all,
Welcome to the group. Yes, this list is primarily a mouthpiece for
my opinions (some say rants) about Radio Userland. It's a great
product, have no doubt, but one with room for improvement. I've not
found what I consider reasonable cooperation from Dave at Userland.
You may find otherwise.
To that end, read the messages here. Take their tone with a grain of
salt. They're not intended to be negative toward the product. Some
of the things I'm seeing new users run up against are old bugs or
missing features. You'll see my posting about them in the message
archive.
So, ask away. This is one place you won't get kicked off for being
opinionated.
Thanks,
Bill Kearney
Well, what an interesting corner of the web!
Anyways, seeing as we have a proprietary (=flexible?) output format,
why not campaign for some user entered metadata. After all, any is
good, yes?
My opening gambit is for an option box for timely/not timely, and a
selection box for a score of 1 to 10.
Oh, and on the systray suggestion - because non-enthusiast users, and
I consider myself an enthusiast here, find the application interface
so intimidating, I think it should dissapear to the systray on
launching. Installation might sprinkle some icons on the desktop to
the web page interface.
I've been looking at the 0.92 rss files that RU produces.
Hmmm? Each <item> has a <description> and nothing else. I've had a run
in with our favourite blogger about this in the past. The subject of
something strange called "Permalinks" always comes up. The more I think
about this, the more it looks like plain old bloody mindedness. Even for
manila based blogs ISTM:-
<title> can be synthesized from the first 25 chars of the item
eg "As today's Microsoft-Free Friday, I downloaded and installed Opera"
<link> could be the date orientated page url for that day only. With
only a little code, it could include a relative link eg
http://roguemoon.manilasites.com/2001/06/15#item23
This would bring U and RU in line with the de facto standard in all
other rss sources that <link> points at the html of the page that
contains the item. And <title> is the headline.
But oh no, "Long time bloggers are more interested in promoting the link
than the words about the link". Or in other words, that's the way we've
always done it and we're not going to change now.
--
Julian Bond email: julian_bond@...
CV: http://www.voidstar.com/cv/
HomeURL: http://www.shockwav.demon.co.uk/
WebLog: http://roguemoon.manilasites.com/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
Julian,
you seem to have made some valid points that could increase the
objectivity of data about data. However, I'm not sure I understand
how what you've written suggests treatment for information about a
link where the link is to an external resource. Would such a link
just form part of <description>?
Given the caveat that I've misunderstood you, I think its less likely
to be a case of *that's the way we've always done it* than *this is
what we do to save time for our readers*.
With my rss consumer hat on: I am intensely irritated by descriptions
that seem to refer to a link the title of which suggests an external
resource, when in fact the link is to the local resource.
Ah, I may have talked my way into understanding you - are you
suggesting that I have to visit your page to access the external
link? If that is the case, why would I want to do that when I have
all the data I need in the feed (except the external url if what I
say in para 1 isn't the case)?
Maybe you do suggest treatment for both local and external links in
the feed, and I just don't get it. Hoping I'm not exasperating,
Charles Goodier (http://doid.com/)
<julian_bond@v...> wrote:
> I've been looking at the 0.92 rss files that RU produces.
> Hmmm? Each <item> has a <description> and nothing else. I've had a
run
> in with our favourite blogger about this in the past. The subject of
> something strange called "Permalinks" always comes up. The more I
think
> about this, the more it looks like plain old bloody mindedness.
Even for
> manila based blogs ISTM:-
>
> <title> can be synthesized from the first 25 chars of the item
> eg "As today's Microsoft-Free Friday, I downloaded and installed
Opera"
>
> <link> could be the date orientated page url for that day only. With
> only a little code, it could include a relative link eg
> http://roguemoon.manilasites.com/2001/06/15#item23
>
> This would bring U and RU in line with the de facto standard in all
> other rss sources that <link> points at the html of the page that
> contains the item. And <title> is the headline.
>
> But oh no, "Long time bloggers are more interested in promoting the
link
> than the words about the link". Or in other words, that's the way
we've
> always done it and we're not going to change now.
>
> --
> Julian Bond email: julian_bond@v...
> CV:
> HomeURL: http://www.shockwav.demon.co.uk/
> WebLog: http://roguemoon.manilasites.com/
> M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
> ICQ:33679668 tag:So many words, so little time
--- In radio-features@y..., Julian Bond <julian_bond@v...> wrote:
> I've been looking at the 0.92 rss files that RU produces.
> Hmmm? Each <item> has a <description> and nothing else. I've had a
> run in with our favourite blogger about this in the past.
...
> But oh no, "Long time bloggers are more interested in promoting the
> link than the words about the link". Or in other words, that's the
> way we've always done it and we're not going to change now.
If third parties aren't unwelcome, after the discussion on the
reallySimpleSyndication list, I made sure my Stapler tool [1] for
Radio could provide a headline-style feed (unfortunately still in
0.92, but that's on my list). 1.1.3 includes a cleaned-up version of
the scraper script I use to make my headline feed [2], including using
the first seventy characters of the item as the title, when one
doesn't use HTML comments as I have been.
[1] http://markpasc.home.mindspring.com/code/stapler/
[2] http://ourfavoritesongs.com/users/
markpasc%40mindspring.com/staplerFeeds/markpasc.xml
Mark Paschal
markpasc@...
In article <9hklib+dmcl@eGroups.com>, doid@... writes
>Maybe you do suggest treatment for both local and external links in
>the feed, and I just don't get it. Hoping I'm not exasperating,
Not at all.
Say I'm Slashdot. An Item looks roughly like this.
<Item>
<title>An Amazing News Story</title>
<link>http://www.slashdot.org/stories/storyid=23</link>
<description>Here's my amazing story with lots of links in
it.</description>
</Item>
So the <Link> refers to the page where you can read the full text to
which the rss item is an abstract. This is the way the vast majority of
feeds are built.
Now consider a blog entry. I've just written
"This <A HREF="http://www.kewl.com">link</A> is amazing you must visit
it." How should this be represented. One possibility is this.
<Item>
<title>This link is amazing</title>
<link>http://myBlog.com/2001/06/15#item23</link>
<description>"This <A HREF="http://www.kewl.com">link</A> is
amazing you must visit it."</description>
</Item>
Or it could be written as
<Item>
<link>http://www.kewl.com</link>
<description>"This <A HREF="http://www.kewl.com">link</A> is
amazing you must visit it."</description>
</Item>
Or as RU does it.
<Item>
<description>"This <A HREF="http://www.kewl.com">link</A> is
amazing you must visit it."</description>
</Item>
But in both cases there is no link to the next paragraph I wrote where I
describe why the link is so amazing. And what happens when I include two
links or more? Which link do I put in <link>?
0.92 has an additional element <source> that "Its value is the name of
the RSS channel that the item came from, derived from its <title>. It
has one required attribute, url, which links to the XMLization of the
source.". I don't understand what this is. ISTM to be something that is
only relevant when you're making a comment about an item in another
feed. But my most recent blog didn't come from and doesn't reference a
feed.
There's something fundamental here that I don't seem to be able to
convey. Is an RSS item metadata about an HTML fragment somewhere on the
web? Or is it that fragment in a form that can be transported somewhere
else? Most of the time it seems to be a headline abstract (metadata).
But Bloggers can listen to a different drummer and use it quite
differently.
<Weeps>Why?</Weeps>
--
Julian Bond email: julian_bond@...
CV: http://www.voidstar.com/cv/
HomeURL: http://www.shockwav.demon.co.uk/
WebLog: http://roguemoon.manilasites.com/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
Ok, thanks. Personal opinions follow...
--- In radio-features@y..., Julian Bond <julian_bond@v...> wrote:
> <Item>
> <title>This link is amazing</title>
> <link>http://myBlog.com/2001/06/15#item23</link>
> <description>"This <A
HREF="http://www.kewl.com">link</A> is
> amazing you must visit it."</description>
> </Item>
On sight, this appeals to me most - transport the whole lot.
> But in both cases there is no link to the next paragraph I wrote
where I
> describe why the link is so amazing. And what happens when I
include two
> links or more? Which link do I put in <link>?
Why can this next paragraph and any subsequent links not be contained
in the <description> element?
> 0.92 has an additional element <source> that "Its value is the name
of
> the RSS channel that the item came from, derived from its <title>.
It
> has one required attribute, url, which links to the XMLization of
the
> source.". I don't understand what this is. ISTM to be something
that is
> only relevant when you're making a comment about an item in another
> feed.
I agree, and if you're trying to comment, why reference an ephemeral
feed? So here's the idea from rss093 [1]
The attribute for this <source> element could be the url for the HTML
fragment that the recent blog references, derived from the <link>
element of the blog entry.
>But my most recent blog didn't come from and doesn't reference a
> feed.
So in this case the most recent blog could be considered the first
message in a thread, a conversation initiator, or "first found" item.
I might then make a blog which references your kewl blog (using
http://myBlog.com/2001/06/15#item23 as the <source> attribute) and,
hey, we have a thread that can branch at any point, where each
posting is syndicated, and has a permanent url.
> There's something fundamental here that I don't seem to be able to
> convey. Is an RSS item metadata about an HTML fragment somewhere on
the
> web? Or is it that fragment in a form that can be transported
somewhere
> else? Most of the time it seems to be a headline abstract
(metadata).
> But Bloggers can listen to a different drummer and use it quite
> differently.
>
> <Weeps>Why?</Weeps>
You've hit a proverbial nail here. It's both, because thats the way
it is.
While I don't feel it, I understand your pain - actually I don't
agree with you that most publishers use the headline abstract method,
just the larger more commercial ones that have advertising on the web
page, and which are now outnumbered (if not in page views) by
individual webloggers who, rightly or wrongly (and maybe as a result
of the tools they use, writing styles and so on), make entire
fragments available for transport.
I prefer transportable fragments (and they are a premise for the
above) as it makes the information freer. I understand that many
consider this heretical as transporting the whole lot is not, in
fact, providing data about data, but this is for some other place.
Just ideas...
Charles Goodier (http://doid.com/)
[1] http://backend.userland.com/rss093
In article <9hntnt+92fo@eGroups.com>, doid@... writes
>Ok, thanks. Personal opinions follow...
>You've hit a proverbial nail here. It's both, because thats the way
>it is.
>
>While I don't feel it, I understand your pain - actually I don't
>agree with you that most publishers use the headline abstract method,
>just the larger more commercial ones that have advertising on the web
>page, and which are now outnumbered (if not in page views) by
>individual webloggers who, rightly or wrongly (and maybe as a result
>of the tools they use, writing styles and so on), make entire
>fragments available for transport.
>
>I prefer transportable fragments (and they are a premise for the
>above) as it makes the information freer. I understand that many
>consider this heretical as transporting the whole lot is not, in
>fact, providing data about data, but this is for some other place.
As for what is the de facto standard; Is Moreover, Newsisfree, 10.am,
Magportal, all the "Slash" sites, and all the commercial sites enough
for you? What about the aggregators like Headline Viewer? What about the
displays of Moreover news that many many people put on their web sites.
This is not RSS, but the data displayed is exactly the same metadata. A
Title, A Link and maybe an abstract. Why should Headline Viewer and
Newsisfree (and others) have to synthesize a title and link from the
blog feeds?
Like yourself, DW's response to this was "display the description". But
maybe I don't want to. Maybe I don't have the web page space.
I get the feeling that this may well be at the core of the 0.92 vs RDFss
1.0 debate.
There is a solution here. For Manila, RU and Userland to produce 2 (or
more) feeds. [1]An RSS 0.91 compatible feed; An RSS 0.92/3 feed for
their own transport; Scripting News format for their own transport.
The 0.91 feed would probably be laid out as I suggested.
Title = 1st 25 chars of the item
Link = Page Archive URL plus item anchor
Description = full item text including escaped html
Now that wasn't hard was it?
The reason I'm so up tight about this is that I'm trying to write code
in PHP that aggregates multiple feeds on a single topic from multiple
sources. I then display this as a composite or by source with and
without the description text, depending on available space. I'm having
to leave out blogs as a source. The alternative is to try and synthesize
the missing elements which is more code and is not going to work very
well.
--
Julian Bond email: julian_bond@...
CV: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/blog/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
Trying to find a solution...
--- In radio-features@y..., Julian Bond <julian_bond@v...> wrote:
> In article <9hntnt+92fo@e...>, doid@d... writes
> There is a solution here. For Manila, RU and Userland to produce 2
(or
> more) feeds. [1]An RSS 0.91 compatible feed; An RSS 0.92/3 feed for
> their own transport; Scripting News format for their own transport.
> The 0.91 feed would probably be laid out as I suggested.
Well, Manila sites do have rss091 feeds at
http://manilasite.ext/xml/rss.xml and Radio websites could use the
Stapler headline feed [1]
But I guess you knew this and would like them both to be the same? Or
is there something else I missed at reallysimplesyndication?
Charles Goodier (http://doid.com/)
[1] http://groups.yahoo.com/group/radio-features/message/76
In article <9hpi4s+ef3g@eGroups.com>, Charles Goodier <doid@...> writes
>Well, Manila sites do have rss091 feeds at
>http://manilasite.ext/xml/rss.xml and Radio websites could use the
>Stapler headline feed [1]
Take this example from
http://scriptingnews.userland.com/xml/rss.xml
- <item>
<title>Luke Tymowski</title>
<link>http://tripping.seeto.com/history/articles/microsoftAlternatives.html</lin\
k>
<description>Luke Tymowski: "I resolved long ago to free myself from
Microsoft's
clutches at home." Excellent piece. The first step towards reinvigorating the
software
industry is to review software. Software developers used to live by the
reviews.</description>
</item>
I would actually only make two small changes to this and make it read:-
- <item>
<title>Luke Tymowski</title>
<link>http://scriptingnews.userland.com/backissues/2001/07/01#02</link>
<description><A
HREF="http://tripping.seeto.com/history/articles/microsoftAlternative
s.html">Luke Tymowski</A>: "I resolved long ago to free myself from Microsoft's
clutches at home." Excellent piece. The first step towards reinvigorating the
software
industry is to review software. Software developers used to live by the
reviews.</description>
</item>
Now if I want to go to the full text of the piece I can with Title plus Link. If
I want
to read it as distributed, I can read the description with the embedded link. It
does
break the 0.91 spec but only by including escaped html which everyone does
anyway.
But compare this with http://www.scripting.com/rss.xml which comes out of RU
direct.
It's 0.92 with no link or title. And it appears that DW doesn't use RU the way I
am any
more as the rss file is dated Jan 2001.
- <item>
<description>NY Times: <a
href="http://www.nytimes.com/2001/01/18/technology/18SELF.h
tml">Web Sites Begin to Self Organize</a>. The Vines is an example of an
emerging class
of what are called self-organizing Web sites. Such sites are demonstrating that
with a
dab or two of well-written code and a bit of careful planning, a site can take a
random
collection of links or posts and turn them into a sophisticated, adaptive
system.</description>
</item>
----------------------------------------------------
Then we have
http://scriptingnews.userland.com/xml/scriptingNews2.xml
- <item>
<text>Luke Tymowski: "I resolved long ago to free myself from Microsoft's
clutches at
home." Excellent piece. The first step towards reinvigorating the software
industry is
to review software. Software developers used to live by the reviews.</text>
- <link>
<url>http://tripping.seeto.com/history/articles/microsoftAlternatives.html</url>
<linetext>Luke Tymowski</linetext>
</link>
</item>
Yet Another XML file and protocol (YAXMLP)! As <scriptingNews> predates 0.91 we
can
begin to see where all this comes from. <url> and <linetext> have been converted
directly to <link> and <title>, and quick and dirty code turned into a virtue.
You see
<scriptingNews> never did have a permalink to the blog page, <irony>so clearly
we don't
need one now</irony>!
And 'jeez, why should I have to use Stapler on Radio Userland! Doesn't this tool
come
from the very home of RSS?
I think I better go and lie down, before I start thinking about cloning Radio
Userland
and Manila in PHP, MySQL and Apache... You see I find the ideas incredibly kewl.
But
I'm getting horribly irritated by the implementation.
--
Julian Bond email: julian_bond@...
CV: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/blog/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
Hi all,
Contrary to Dave's latest comments, Radio Userland is still a beta
product. It has tons of things that a release version shouldn't
possess. Read the archives here and you'll understand.
The ftp features are bug riddled. The uploading to cloud, ftp,
static and local are very inconsistent. The handling of archives is
in need of serious rearchitecting.
In short, it's a great beta product. I look forward to it's
refinement into something worthy of being called a release.
-Bill Kearney
In article <9hscfj+3bfq@eGroups.com>, Bill Kearney
<wkearney99@...> writes
>Contrary to Dave's latest comments, Radio Userland is still a beta
>product. It has tons of things that a release version shouldn't
>possess. Read the archives here and you'll understand.
>
>The ftp features are bug riddled. The uploading to cloud, ftp,
>static and local are very inconsistent. The handling of archives is
>in need of serious rearchitecting.
>
>In short, it's a great beta product. I look forward to it's
>refinement into something worthy of being called a release.
Indeed.
I've got so frustrated with manilasites and the Radio upload that today
I converted to ftp uploading to a personal hosted site. As I have
complete control of this, I've mixed a PHP display system with Radio's
blogging. You can see the results at http://www.voidstar.com
The only Radio contributed bit is the center text section. Everything
else, including the calendar, comes from php.
One of the things that still winds me up is that if you use RU, you have
to stop using Manila. Any text entered in Manila is wiped next time you
post with RU.
Anyway, this is an interim step while I think about how complex
completely replacing Manila and Radio with PHP would be. I'm seriously
tempted but it's going to be quite a lot of code, even starting with the
Drupal 'Slash' site environment. The other pieces leading me down this
route are http://www.b2bslash.com and http://www.bikeweb.com both of
which I've built in the last few months. ISTM that there are three
strands to this puzzle.
1) Slashdot, Kuro5hin community journalism sites.
2) Single user personal blogs with complete control over code, data and
environment
3) Hosted versions of 2) where lots of people can build slightly
restricted versions with templated UI.
All three can benefit from some of Radio's ideas about RSS display,
editing and routing. But the whole Radio environment feels like a sledge
hammer to crack the nut. And a sledge hammer that was originally
designed to be a screwdriver.
ps. Still looking for a job. But then it's not yet a week since the last
one ended.
--
Julian Bond email: julian_bond@...
CV: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
> One of the things that still winds me up is that if you use RU, you
have
> to stop using Manila. Any text entered in Manila is wiped next time
you
> post with RU.
This is what I find so appalling about how Userland has handled
Radio. I mean, get real, this is an important area that deserves
attention. But noooooo...
Thus it takes the initial idea and forces people to run ELSEWHERE for
solutions. That and Dave's own personality seems hell bent on
driving away the people that are capable of helping.
> Anyway, this is an interim step while I think about how complex
> completely replacing Manila and Radio with PHP would be. I'm
seriously
> tempted but it's going to be quite a lot of code
It is and it isn't. I too have entertained this question. It's
really not as hard as it seems. However, if you want to take it's
features, FIX it's bugs and develop new features, it is going to take
some coding.
> ISTM that there are three strands to this puzzle.
>
> 1) Slashdot, Kuro5hin community journalism sites.
> 2) Single user personal blogs with complete control
> 3) Hosted versions of 2)
>
> All three can benefit from some of Radio's ideas about RSS display,
> editing and routing.
Yes, I agree. That ROUTING is an important area that's only barely
implemented in Radio. Have you seen Conversant from Macrobyte?
> But the whole Radio environment feels like a sledge
> hammer to crack the nut. And a sledge hammer that was originally
> designed to be a screwdriver.
Yes again. If Dave would only listen to the idea of making an
updated version of MORE that understood outlining and the web.
Instead he's made this almost-Frontier program. It's a crappy
outliner (compared to others like) and it's tools are in need of
serious improvement.
So it seems the consensus is to use or develop something else.
-Bill Kearney
I spent the morning messing around in the bowels of Radio as I don't
like the way it generates 0.92 RSS files and the <channel>.<link>
element wasn't how I wanted it. I've ended up (for now) with <link> tags
on all the items that point to <A Name>s on the correct archive page.
Along the way I discovered that Radio has no concept of an item ID. It
just iterates over all of them. This may be in part a cause of DW's
aversion to "PermaLinks". Radio simply doesn't store the data in a form
where these could be created. As someone who comes from a database
background I find this a little shocking. The first thing you do on
almost any table is to add an Auto-Increment, Unique primary key. But
then I've seen the same mistake being made on other object databases.
As it is, I really don't like the development environment one bit. I
know everybody has their own favourite which typically inspires
religious fervour. But at least with most scripting languages you get
some indication when things are going wrong. Radio has a nasty habit of
just giving up without any outside indication. Makes debugging, a pig.
I'm really not sure why I keep persisting with this. It occurs to me
that I could have written a few hundred lines of PHP in the time I spent
making two smallish adjustments. And in the knowledge that when I next
update Radio they'll get wiped out. Maybe even in an automatic update
while I'm not looking.
Hey-Ho.
--
Julian Bond email: julian_bond@...
CV/Resume: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
--- In radio-features@y..., Julian Bond <julian_bond@v...> wrote:
> I spent the morning messing around in the bowels of Radio as I don't
> like the way it generates 0.92 RSS files and the <channel>.<link>
> element wasn't how I wanted it. I've ended up (for now) with <link>
tags
> on all the items that point to <A Name>s on the correct archive
page.
>
> Along the way I discovered that Radio has no concept of an item ID.
It
> just iterates over all of them. This may be in part a cause of DW's
> aversion to "PermaLinks". Radio simply doesn't store the data in a
form
> where these could be created. As someone who comes from a database
> background I find this a little shocking. The first thing you do on
> almost any table is to add an Auto-Increment, Unique primary key.
But
> then I've seen the same mistake being made on other object
databases.
>
> As it is, I really don't like the development environment one bit. I
> know everybody has their own favourite which typically inspires
> religious fervour. But at least with most scripting languages you
get
> some indication when things are going wrong. Radio has a nasty
habit of
> just giving up without any outside indication. Makes debugging, a
pig.
>
> I'm really not sure why I keep persisting with this. It occurs to me
> that I could have written a few hundred lines of PHP in the time I
spent
> making two smallish adjustments. And in the knowledge that when I
next
> update Radio they'll get wiped out. Maybe even in an automatic
update
> while I'm not looking.
Yep, you're treading into the same morass I ran up against last
winter. What a mess. Cool stuff but seemingly just half-
implemented. This, I suppose, is what happens when you refuse to
listen to other people...
But Julian, you really have to stop trying to run it in Windows ME.
It just doesn't work. It even tends to blue screen Windows 2000 now
and then. I don't let it run for more than 4 days straight. Also,
be sure you're purging stories to the archives. This has a great
impact on performance/stablity of Radio. I've had it collect over
6000 new 'items' and then take over a HALF HOUR for the application
to do an orderly quit!! Memory management seems sorely lacking in
Radio.
Don't get me started on debugging.
And yes, isn't it supremely stupid for Radio (and Frontier) to have
no way to support source/version changes?
-Bill Kearney
--- In radio-features@y..., Julian Bond <julian_bond@v...> wrote:
> Along the way I discovered that Radio has no concept of an item ID.
> It just iterates over all of them. This may be in part a cause of
> DW's aversion to "PermaLinks". Radio simply doesn't store the data
> in a form where these could be created.
For the Radio-published weblog? The numbers with which the tables in
myUserLandData.blogs.default.posts are named seem to be passable item
IDs, whether MUOTD uses them or not. Personally I've had good success
using "string.dropNonAlphas (date.timeString (item^.when))" for a
permalink anchor on archive pages.
> I'm really not sure why I keep persisting with this. It occurs to me
> that I could have written a few hundred lines of PHP in the time I
> spent making two smallish adjustments.
I would love to have a MUOTD (emphasis on "Desktop," for me) that
happened to be easier to change and improve. The worst part about
MUOTD is it needs Radio to run.
Is anyone working on aggregator-weblog combination software, other
than Peerkat? (I ask for scholarly research type purposes, of course.)
> And in the knowledge that when I next update Radio they'll get
> wiped out. Maybe even in an automatic update while I'm not looking.
I haven't had any changes to myUserLand.root wiped out yet, but that's
because (as our friend Dave has said) they know people are messing
with it, and they're working on "Smurf Turf" or whatever anyway. Not
that that means I haven't kept backup copies of my changes in
plaintext files elsewhere on disk--some kind of versioning like Bill
says would be great.
Mark Paschal
markpasc@...
In article <9i7qdt+bftg@eGroups.com>, Mark Paschal
<markpasc@...> writes
>I would love to have a MUOTD (emphasis on "Desktop," for me) that
>happened to be easier to change and improve. The worst part about
>MUOTD is it needs Radio to run.
>
>Is anyone working on aggregator-weblog combination software, other
>than Peerkat? (I ask for scholarly research type purposes, of course.)
I've asked elsewhere if there is any open source code out there for
Blogger/Manila style Personal blogs or clouds of such. We didn't find
much. But there's loads of it for SlashDot style community journalism.
This is all we found:-
http://www.free-conversant.com/http://www.livejournal.com/http://livejournal.com/code/
So with that in mind the guys behind Drupal (http://www.drupal.org) and
I have begun to think about how to architect such a thing. Drupal V3 is
mostly a SlashCode type of thing, but written in a very modular way and
with some Wiki like community authorship functions. It already has a lot
of the bits in place including RSS feed aggregation into bundles. And
I've done some work already on using feed items as the source of new
stories like the "post" buttons on the Radio home page.
So don't hold your breath, but...
--
Julian Bond email: julian_bond@...
CV/Resume: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
We've just added a News headline and Blog system to the development copy
of Drupal. This replicates some of the function of Radio and Manila in
these areas. More detail at http://www.voidstar.com/
Drupal is a highly modular, GPL, portal system written in PHP and MySQL.
It can be used as a local system with Apache and MySQL locally, but it's
main purpose is to build Slash Style sites. It's broadly similar to
Slashcode, Scoop and PHP-Nuke. More detail at http://www.drop.org
--
Julian Bond email: julian_bond@...
CV/Resume: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
This is interesting news:
http://support.microsoft.com/support/kb/articles/Q292/4/42.asp?
SD=MSDN&FR=0
I mentioned this a while back. When will Radio stop reinventing the
wheel and actually USE the very powerful features provided already by
the operating system?
-Bill Kearney
I faced the same dilemma of how to use Radio to publish to an
FTP site, and offer all the other navigational elements that make
up a site, but are not part of the blog content nor would you want
it rendered each time.
Possible workaround I looked at incude:
Loading a rendered weblog page into a url or iframe. - Works
quite well on modern browsers, but is slow.
Using SSI's, if you have that available.
Using macros. Requires lots of understanding of radio
Adding HTML to your categories description, then cleverly placing
the {categories} tag in your page layout.
Using Frames.
All these possiblities have left me with the impression that there
is no practical need for Manilla, and it's better to go the FTP
route.
Granted the FTP has hardly been reliable for me, and I just use
Dreamweavers FTP instead, but overall, after a fortnight of
tinkering, I've managed to get radio nicely integrated into my site.
http://home.iprimus.com.au/joris/
> One of the things that still winds me up is that if you use RU,
you have
> to stop using Manila. Any text entered in Manila is wiped next
time you
> post with RU.
>
Brent Simmons <brent@...> wrote:
>Thanks for the reports, Julian.
>
>2. You're correct that Manila never returns a 404.
Yup, Manila always returns "200 OK" no matter what happened.
Feature request. Manila should return all the normal error codes just
like every other web server.
--
Julian Bond email: julian_bond@...
CV/Resume: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679668 tag:So many words, so little time
I have created a new resource site for Radio UserLand users. The site
includes Radio UserLand resources and Frontier material that could
help Radio UserLand users. Check it out at:
http://ruminations.weblogger.com
Andy Sylvester
ruminations@...
Ruminations - news and resources for Radio UserLand users
http://ruminations.weblogger.com
--- In radio-features@y..., ruminations@c... wrote:
> I have created a new resource site for Radio UserLand users. The
site
> includes Radio UserLand resources and Frontier material that could
> help Radio UserLand users. Check it out at:
>
> http://ruminations.weblogger.com
Thanks for the link!
-Bill Kearney
It would be handy if there was a way to tell when a scan had taken place while
looking at the main view. Perhaps nothing more than a small line of text with
the date, time, number of channels and total number of new stories on it.
Essentially the same thing as shown in the events list. Yes, this is one click
away but not if you're on a mobile device that's not downloading that page.
-Bill Kearney
[Non-text portions of this message have been removed]
I'm getting increasingly pissed about Manila's inability to support RSS
correctly. After posting bug reports in 3 or 4 different places with
zero reaction, I started chipping in on the Syndication list (as some of
you know). The reaction from Dave (aka "Cartman") was "If you lecture me
in public, don't expect me to co-operate". So I'm working on the
assumption that he is now deliberately refusing to fix any problems in
Manila's RSS out of spite.
I've almost come to the conclusion that I will simply unsubscribe from
RSS feeds from Manila and stop reading any text from any site that is
running on that platform. I've already done this with Doc Searles as his
output is simply unreadable. This is really sad as there is plenty of
good content being generated that is worth reading. Another ironic data
point, Adam Curry (http://www.curry.com) has stopped pointing to the
Manila RSS on his site for the same reasons and points instead to the
Radio Ourfavorite RSS which is marginally better.
So as a final lob, I sent this to Dave, copy John Robb. If this is
affecting you as well, I'd suggest sending a similar email to them and
maybe to the programmers Brent and Jake as well. Though I don't hold out
much hope of it changing anything.
----------------------------------------------------------------------
Bug report
For the record, here's a list of issues I'm aware of with the rss file
generated by Manila. This is the file typically located at /xml/rss.xml
I appreciate that some of these may be awkward to change, or may take
time to change that might be better spent on other things. My
suggestions may also not be the best way of solving them. I'm not
demanding that "you fix this now". I'm simply making sure that you're
aware of the issues that your users are facing.
Bugs
1. Somewhere in the code that strips <a Href> tags from the text,
is a function that is incorrectly removing them. It is failing
to remove sub elements of <A> such as title. One of the symptoms is that
[<A Href="url" title="Permanent link to this item in archive.">] gets
converted to [title="Permanent link to this item in archive.">] With the
appearance of the new permalink function, this is getting extremely
annoying as every <description> of every <item> now ends with this text.
Further, since the link html has been removed, the permalink only
appears in an <A name> tag in the RSS and then only the anchor part of
it.
2. Usually related to 1. Manila is leaving html in the <title> element
of <item>s. This is frequently an <img> tag. This breaks all of the 0.9x
specs.
3. Manila occasionally fails to correctly encode XML entities. There's
an example in the current scripting news where a high order apostrophe
in "Israel's" is not encoded. This is breaking IE5's display which may
be Microsoft's problem, not yours. This is definitely not critical.
Annoyances
1. The Manila RSS at /xml/rss.xml is breaking 0.91 by leaving embedded
html in the <description>. This is common and not a problem. However it
is stripping all the <a href> tags. It would be considerably more useful
if these were left in place. This would preserve the embedded links
which are a common feature of blog content. As noted above, the code
that strips <a href> is broken but it would be better if it didn't strip
them at all.
2. The RSS file only includes items for the current day, not the last N
items. This has an implication for the reader that they must collect the
rss file between the last update and the next page flip or miss the
content. It would be more useful if it contained something like current
day plus N items, or just last N items.
3. There is a way of updating the text (probably by using Radio as an
outline for the whole day) that leaves Manila thinking that a whole
day's text is a single item when it comes to generate the rss. This is
despite the content being broken into multiple sections at some stage in
the process. It would be more useful to RSS readers if it was broken
into items at the first sub-level of the opml hierarchy.
4. It's unfortunate that the new permalink function doesn't generate a
true permalink but rather one that changes with each edit. Since it's
only the piece to the right of the "#" that may change, this is an
annoyance rather than a bug, since the permalink will still point at the
right page. However, the existence of the permalink now makes it
possible for Manila to generate <link> elements in the RSS that point at
the permalink rather than one of the links in the content. This would
bring Manila in line with the common practice of most other RSS
generators.
Hope this helps. No reply needed.
--
Julian Bond email: julian_bond@...
CV/Resume: http://www.voidstar.com/cv/
WebLog: http://www.voidstar.com/
HomeURL: http://www.shockwav.demon.co.uk/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433
ICQ:33679568 tag:So many words, so little time
> 4. It's unfortunate that the new permalink function doesn't
generate a
> true permalink but rather one that changes with each edit. Since
it's
> only the piece to the right of the "#" that may change, this is an
> annoyance rather than a bug, since the permalink will still point
at the
> right page. However, the existence of the permalink now makes it
> possible for Manila to generate <link> elements in the RSS that
point at
> the permalink rather than one of the links in the content. This
would
> bring Manila in line with the common practice of most other RSS
> generators.
At the very least it would help if they used a PermaLink with meta-
data that allowed a client package to differentiate between the
original and the currently online version of the item. Using an MD5
hash as the permalink is less than ideal. Using the MD5 hash in
conjunction with a genuine per-item permalink has possiblities.
And you're right, the arrogance of the folks at UserLand in how they
selectively exclude people from their conversations is staggering.
It's fair to say they've pissed away a tremendous amount of support
they could have been getting. So instead of taking their lumps
gracefully and reaping the benefits, they've pissed people off and
they LEFT. While Dave might think he's the greatest thing since
sliced bread, he'll never really know. For all the people that
question his perception of reality get tossed aside. It's easy to
think you're so smart when only the dumb people listen to you anymore.
-Bill Kearney
Hi all,
I've been thinking more about categories and actions related to
them. I'm certainly in favor of a way for Radio to export it's
categories to an XML file. This so others can use the categories on
other sites or as a means to traverse other meta-data.
When using categories I'm also interested in being able to trigger
actions based on those categories. The folks writing Conversant have
some cool ideas about using mailing lists and NNTP interfaces to
blogs.
I'm thinking more along the lines of being able to flag an item right
as I'm blogging it. I'd really like to be able to grab a chunk of
info, put some categories on it, prepare to blog it AND have it allow
me to do a little more post-blog. To that end, I'm going to
experiment a bit with putting some extra checkboxes on the blogging
screen that are tied to actions.
My first thoughts about actions surround being able to mail a blogged
item. This way I can grab something, categorize it away in my own
table also mail it to someone else AFTER I hit the Post button. I
realize I could just go back to the blogged item and add something on
that screen that says 'mail this page' or whatever. But that's not
how to the thought process works. I see something, I know someone
else that would be interested in it. I want to keep my own copy and
send one to them. This should be a FLOW oriented process, not the
jumping all around dance we have to do now.
This leads to thoughts about using a mail list system; aka Listserv.
In order to get the word out to more people it would be good to have
the categories understand how to send messages to a mailing list. If
not manage a maillist list inside of Radio, at least use one on the
outside. This is not a trivial undertaking. Running a mail list is
a bit of an adventure, mainly because membership management. Using
one mail list per category is what's necessary. The hassle being
management of the membership. You wouldn't want to force everyone to
sign up to each and every list manually. Sure, if they liked your
material they'd do it. But it's more tedious than most folks want.
Thus a user interface that allowed for checking off which categories
they'd like to receive is probably the best idea.
Of course, an alternative would be to put the categories into some
kind of header meta-data. This breaks down, however, because most
mail clients don't really handle per-message meta-data properly.
I think using category meta-data in headers is a good idea. Do any
of the RFCs cover this concept?
This leads to NNTP integration. Why not have categories associated
with NNTP readable newsgroups. I'm a BIG fan of NNTP mainly because
of the number of clients and their features. I can read a lot more
stuff via a fast NNTP client than any web-based functions. The
sorting alone is a tremendous improvement. And the NNTP clients
generally understand decent printing and mail forwarding functions.
Again, putting an NNTP or mail list server inside of Radio is not a
trivial task. The way Radio handles TCP connections is just took
lame to make it work reliably. But integration with external
products might be worth considering.
So, feedback?
-Bill Kearney
Hi all,
It would be greatly appreciated if the subscriptions.wsf page showed
a date for when the last item was ADDED to a subscribed channel.
Right now it just shows when it last checked the feed. This is good
on one level, in that it shows whether a feed is still online. But
it does nothing about informing you as to whether there's anything
still coming from that feed.
So add a second column that shows date of last added item.
-Bill Kearney
Send Email
