Today I made the mistake of looking at the Publican produced release notes with Internet Explorer. I can't believe it has taken me so long to make this check, but it wasn't pleasant. They are pretty badly munged up. I wonder if there are things we can do/set in Publican to straighten this out.
The first problem is that the first page stops with the logo image. This eliminates the table of contents and first section. By changing the chunking in Publican we can get the first section back, but I don't know how to get the table of contents. And the little box where the logo belongs looks trashy. I suspect I could modify the branding to use the png version. I also did a different document with Publican, and all looked fine when I looked at it on my local box, but when I moved it to a webserver off my LAN (I believe it is RHEL), a bunch of XML shows up where the logo belongs when viewed with FF. And it has the same issues with IE.
Second, every section number is followed by an A circumflex. I suspect this is some sort of codepage issue, but I can't say I'm sure of that.
I checked the Fedora 10 release notes, which weren't produced with Publican, and they are fine.
I know, why aren't I using Firerox/KonquerorOpera pick your favorite poison. Fact is, Ryan put together the F11 RN's months ago, and amongst many experiments, I've probably viewed them a hundred times, with Firefox, Opera, SeaMonkey, Konqueror, Epiphany; I can't believe it has taken me this long to look at them with IE. In spite of the inroads Firefox in particular has made, Internet Explorer is still the big bear in the woods, and it doesn't look good for Fedora if our documentation looks trashy.
--McD
Two things...
1) How does the Security Guide[1] look using IE?
2) Do we know if the guides are W3C validated? I'll go ahead and answer my #2. I just checked the Security Guide[1] against the W3C Markup Validation Service and received "This document was successfully checked as XHTML 1.0 Strict!". That's right boys and girls, all the code is completely within the spec. I didn't even get a warning or anything. Publican is pushing good html. Which leads to the question, what version of IE are you using?
Eric
[1] http://docs.fedoraproject.org/security-guide/f10/en_US/
On Mon, 2009-03-30 at 08:39 -0400, John J. McDonough wrote:
Today I made the mistake of looking at the Publican produced release notes with Internet Explorer. I can't believe it has taken me so long to make this check, but it wasn't pleasant. They are pretty badly munged up. I wonder if there are things we can do/set in Publican to straighten this out.
The first problem is that the first page stops with the logo image. This eliminates the table of contents and first section. By changing the chunking in Publican we can get the first section back, but I don't know how to get the table of contents. And the little box where the logo belongs looks trashy. I suspect I could modify the branding to use the png version. I also did a different document with Publican, and all looked fine when I looked at it on my local box, but when I moved it to a webserver off my LAN (I believe it is RHEL), a bunch of XML shows up where the logo belongs when viewed with FF. And it has the same issues with IE.
Second, every section number is followed by an A circumflex. I suspect this is some sort of codepage issue, but I can't say I'm sure of that.
I checked the Fedora 10 release notes, which weren't produced with Publican, and they are fine.
I know, why aren't I using Firerox/KonquerorOpera pick your favorite poison. Fact is, Ryan put together the F11 RN's months ago, and amongst many experiments, I've probably viewed them a hundred times, with Firefox, Opera, SeaMonkey, Konqueror, Epiphany; I can't believe it has taken me this long to look at them with IE. In spite of the inroads Firefox in particular has made, Internet Explorer is still the big bear in the woods, and it doesn't look good for Fedora if our documentation looks trashy.
--McD
----- Original Message ----- From: "Eric Christensen" eric@christensenplace.us To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 9:26 AM Subject: Re: More Publican Pain
I'm using IE7. I believe IE8 is still in beta. Unfortunately, I recently updated all my IE6's. That is still probably the most prevalent browser out there and I can't test it. IE7 appeared about the same time as Vista and it got caught up in the Vista negativity. I haven't seen any evidence of the IE7 horror stories I heard when it first came out, but they prevented me from getting rid of IE6 for a very long time. And on those web pages where I see those kind of stats, I still see a lot of IE6.
I've never seen svg graphics used on the web except on Fedora sites, but it is just plain wrong for IE to simply barf on that. I would expect to see the box and go on. Why it totally stops I have no idea.
On the other hand, I have no idea why Publican would put out a 0xc2 after each section number. This seems like an unnecessary thing to do.
*HOWEVER*, if I select UTF-8 encoding, the page displays properly. The A circumflex is clearly an IE bug. With encoding set to automatic, IE selects Western European (at least for a US copy of Windows), when the page clearly says UTF-8.
Clearly, there are a couple of IE bugs here that we are exploiting. But it really doesn't change the problem, does it?
--McD
On Mon, 2009-03-30 at 09:56 -0400, John J. McDonough wrote:
----- Original Message ----- From: "Eric Christensen" eric@christensenplace.us To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 9:26 AM Subject: Re: More Publican Pain
Clearly, there are a couple of IE bugs here that we are exploiting. But it really doesn't change the problem, does it?
--McD
As far as I can tell, it is a problem with IE and not Publican. When the W3C validation comes up clean it is meeting the spec. IE has always had problems meeting the spec, which is unfortunate and not our problem, in my opinion. We are doing our due diligence to make sure it meets the international spec. Microsoft should be doing the same.
Eric
----- Original Message ----- From: "Eric Christensen" eric@christensenplace.us To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 10:31 AM Subject: Re: More Publican Pain
As far as I can tell, it is a problem with IE and not Publican. When the W3C validation comes up clean it is meeting the spec. IE has always had problems meeting the spec, which is unfortunate and not our problem, in my opinion. We are doing our due diligence to make sure it meets the international spec. Microsoft should be doing the same.
I agree that the problem is in IE, but it *IS* our problem.
These issues are livable to the faithful. A little trashy but the important information is all readable.
But the overwhelming majority of those we want to convert will be looking at our docs using the world's most popular browser, seeing the docs are trash, and deciding that Fedora must be trash, too. That, sir, is very much OUR problem.
--McD
On Mon, 2009-03-30 at 11:50 -0400, John J. McDonough wrote:
----- Original Message ----- From: "Eric Christensen" eric@christensenplace.us To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 10:31 AM Subject: Re: More Publican Pain
As far as I can tell, it is a problem with IE and not Publican. When the W3C validation comes up clean it is meeting the spec. IE has always had problems meeting the spec, which is unfortunate and not our problem, in my opinion. We are doing our due diligence to make sure it meets the international spec. Microsoft should be doing the same.
I agree that the problem is in IE, but it *IS* our problem.
These issues are livable to the faithful. A little trashy but the important information is all readable.
But the overwhelming majority of those we want to convert will be looking at our docs using the world's most popular browser, seeing the docs are trash, and deciding that Fedora must be trash, too. That, sir, is very much OUR problem.
--McD
Okay, I'm impressed...
So check out the stats[1] for docs.fp.o. While most of our visitors are running Windows (53.4%), most are NOT running IE (only 20.9% versus Firefox users at 63.8%). Still, 1,361,844 people running IE (just this year) is a nice chunk of users.
If we can "fix" the problems without breaking the validation then go for it, but we should be meeting an international standard and not what Microsoft feels like using for their browser.
Can we file a bug against IE7?
Eric
[1] http://fedoraproject.org/awstats/docs.fedoraproject.org/
----- Original Message ----- From: "Eric Christensen" eric@christensenplace.us To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 12:22 PM Subject: Re: More Publican Pain
If we can "fix" the problems without breaking the validation then go for it, but we should be meeting an international standard and not what Microsoft feels like using for their browser.
Well, neither of these problems seem to have anything to do with validation. The Fedora branding for Publican has both svg and png versions of the logo, I'm sure IE can handle the png. I'm not sure how to tell it to use the png, tho. The logo is embedded as an object, not an image, and I think the validation suite lets you put anything into an object type as long as it is lower case. I suspect Windows can't decode an svg whatever the browser. But in any case, it should leave a box for the object and go on, not just stop displaying the page there.
The 0xc2 is another issue. While I am sure that the validation doesn't require a 0xc2 after a number, why it is there and how to make it go away I have no clue. Apparently it must be a non-display character in UTF. It precedes a non-break space, so maybe it has something to do with that.
Can we file a bug against IE7?
I'm sure we can, I'll look into it. But I don't get the impression M$ treats bugs like we do. We encourage people to file bugs; we almost like them because it means people care enough to help us make the product better. I don't get that same feeling from Microsoft, but then maybe I'm just prejudiced.
My wife's box has Firefox and Opera on it. I tried to fire up Firefox to take a look on FF/Win. But the silly thing kept crashing, and the phone rang before I could investigate the crash.
--McD
On Mon, 2009-03-30 at 12:58 -0400, John J. McDonough wrote:
The 0xc2 is another issue. While I am sure that the validation doesn't require a 0xc2 after a number, why it is there and how to make it go away I have no clue. Apparently it must be a non-display character in UTF. It precedes a non-break space, so maybe it has something to do with that.
0xc2 is the first byte of a 2-byte UTF-8 sequence for a character between U+0080 and U+00BF.
----- Original Message ----- From: "Ignacio Vazquez-Abrams" ivazqueznet@gmail.com To: "For participants of the Documentation Project" fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 3:51 PM Subject: Re: More Publican Pain
0xc2 is the first byte of a 2-byte UTF-8 sequence for a character between U+0080 and U+00BF.
Ahh, that makes sense then. It is always followed by a non-breaking space, 0xa0. For Western European coding I don't think there are any multi-byte sequences. So the problem is entirely on the codepage.
--McD
On Mon, 2009-03-30 at 16:26 -0400, John J. McDonough wrote:
For Western European coding I don't think there are any multi-byte sequences.
The UTF-8 byte sequence for every single Latin-1 character outside of ASCII starts with either 0xc2 or 0xc3, and is two bytes long.
On Mon, Mar 30, 2009 at 08:39:37AM -0400, John J. McDonough wrote:
Today I made the mistake of looking at the Publican produced release notes with Internet Explorer. I can't believe it has taken me so long to make this check, but it wasn't pleasant. They are pretty badly munged up. I wonder if there are things we can do/set in Publican to straighten this out.
The first problem is that the first page stops with the logo image. This eliminates the table of contents and first section. By changing the chunking in Publican we can get the first section back, but I don't know how to get the table of contents. And the little box where the logo belongs looks trashy. I suspect I could modify the branding to use the png version. I also did a different document with Publican, and all looked fine when I looked at it on my local box, but when I moved it to a webserver off my LAN (I believe it is RHEL), a bunch of XML shows up where the logo belongs when viewed with FF. And it has the same issues with IE.
That is a server-side setting. It is not pushing the proper MIME type for SVG. We had this on the docs.fedoraproject.org servers the first time we put up the SELinux Guide; and it wasn't universal, only some servers had the MIME type wrong, so it was harder to figure out.
I'm not sure about the chunking stuff; let's see how it looks once you fix the MIME type.
Second, every section number is followed by an A circumflex. I suspect this is some sort of codepage issue, but I can't say I'm sure of that.
I think is also server-side, it's pushing the wrong character set. Can you confirm the server is set to use UTF-8? It is likely using ISO 8859, which was Apache default for a while (still?)
I checked the Fedora 10 release notes, which weren't produced with Publican, and they are fine.
Are you also seeing the character problems on a local copy or just from the server?
- Karsten
----- Original Message ----- From: "Karsten Wade" kwade@redhat.com To: fedora-docs-list@redhat.com Sent: Monday, March 30, 2009 12:57 PM Subject: Re: More Publican Pain
I'm not sure about the chunking stuff; let's see how it looks once you fix the MIME type.
I'm not sure I can. This is a commercial server, I don't recall seeing anything on MIME types on the control panel.
Can you confirm the server is set to use UTF-8? It is likely using ISO 8859, which was Apache default for a while (still?)
This is the case whether it comes off the commercial server or read as a file over Samba. The header clearly says it is UTF-8, I guess I would expect IE to pay attention to that (although I haven't installed Fiddler to see whether it is getting some other advice). And when I select UTF-8 from IE's menu the weird characters go away.
Are you also seeing the character problems on a local copy or just from the server?
I only checked the F10 release notes from d.fp.o. But, since I have the F10 notes spinning it is a 10 second job to look on a local copy, and they look just fine, except of course, for a different css.
--McD
Wondering who is working on this. Seems out of date since the last version I can access is the draft for Fedora 9.
Whoever is working on this please let me know how I can help. Suggest working on the guide for Fedora 11 first instead of playing catch up. Since I'm new to Linux/Fedora it would also give me a chance to play with it to my heart's content ... as soon as I get to download the beta which is supposed to be available tomorrow.
Wolf
_________________________________________________________________ Chat with the whole group, and bring everyone together. http://go.microsoft.com/?linkid=9650735
Wolf DreamWalker wrote:
Wondering who is working on this. Seems out of date since the last version I can access is the draft for Fedora 9.
Whoever is working on this please let me know how I can help. Suggest working on the guide for Fedora 11 first instead of playing catch up. Since I'm new to Linux/Fedora it would also give me a chance to play with it to my heart's content ... as soon as I get to download the beta which is supposed to be available tomorrow.
Wolf
Windows Live Messenger makes it easier to stay in touch - learn how! http://go.microsoft.com/?linkid=9650731
Welcome Wolf,
We're finishing up the F10 User Guide and will be starting on F11 soon.
Suggest you get on the IRC server freeNode, #fedora-docs, and meet everyone. The lead for the User Guide is danielsm1 on IRC.
-Kirk
On Mar 30, 2009, at 2:56 PM, Wolf DreamWalker wrote:
Wondering who is working on this. Seems out of date since the last version I can access is the draft for Fedora 9.
The best place to find it right now is on the tasks page [1]. I know it seems like User_Guide makes more sense, but since we only use the wiki for editing and collaboration and not for publishing, appearance doesn't matter too much, and User_Guide_tasks already had a task table set up when I took over the lead for the guide.
On Mar 30, 2009, at 5:17 PM, Kirk wrote:
We're finishing up the F10 User Guide and will be starting on F11 soon.
Reiterating what Kirk said, the F10 guide is basically done, content wise. I'm going to grab the MediaWiki formatting off the wiki so that I can convert it to DocBook XML when I get the chance (our MediaWiki -
DocBook conversion tool is in the process of being fixed right
now), and as soon as I do that we can start working on the F11 guide right in the wiki. The F11 final release is scheduled for May 26 with the release candidate about a month before, so we may be able to push an F11 guide out the window by the time the final release is out.
On Mar 30, 2009, at 2:56 PM, Wolf DreamWalker wrote:
Since I'm new to Linux/Fedora it would also give me a chance to play with it to my heart's content ... as soon as I get to download the beta which is supposed to be available tomorrow.
I would suggest you go ahead and download/install the beta since that's our upcoming task. In the meantime, you may want to read through the current F10 draft to get a feel for the writing style and conventions we try to use. It's definitely not perfect though, so if you want to find the final word on writing conventions, look around the style guide on the Docs Team page [1] or ask someone knowledgeable in IRC.
On Mar 30, 2009, at 5:17 PM, Kirk wrote:
Suggest you get on the IRC server freeNode, #fedora-docs, and meet everyone. The lead for the User Guide is danielsm1 on IRC.
I usually have IRC clients open on different shells, so sometimes my nick is danielsm1. If you can't find that, I'm probably danielsmw, which is what I try to use primarily.
Cheers,
Matthew Daniels
Links: [1] - https://fedoraproject.org/wiki/User_Guide_tasks [2] - http://fedoraproject.org/wiki/DocsProject
On Mon, Mar 30, 2009 at 05:34:10PM -0400, Matthew Daniels wrote:
I'm going to grab the MediaWiki formatting off the wiki so that I can convert it to DocBook XML when I get the chance (our MediaWiki -> DocBook conversion tool is in the process of being fixed right now), and as soon as I do that we can start working on the F11 guide right in the wiki.
One idea that is a bit of a PITA but maybe worth it ...
We could "branch" in the wiki by copying all the current pages to e.g. [F11 User Guide - *]]. I think we should not rename because it creates redirects we don't want. We'll also want the F10 content in place to rewrite from.
Just an idea ...
- Karsten
We could "branch" in the wiki by copying all the current pages to e.g. [F11 User Guide - *]]. I think we should not rename because it creates redirects we don't want. We'll also want the F10 content in place to rewrite from.
As long as that doesn't create any wiki problems, that sounds like a good idea. Would we create a new category too, like we did for F8, or should we just leave them both with [[Category:User Guide]]?
- Matthew
On Sun, Apr 5, 2009 at 12:18 PM, Matthew Daniels danielsmw@gmail.com wrote:
We could "branch" in the wiki by copying all the current pages to e.g. [F11 User Guide - *]]. I think we should not rename because it creates redirects we don't want. We'll also want the F10 content in place to rewrite from.
As long as that doesn't create any wiki problems, that sounds like a good idea. Would we create a new category too, like we did for F8, or should we just leave them both with [[Category:User Guide]]?
New Category. [[Category:F11 User Guide]] Though you can have [[Category:User Guide]] with subcats for [[Category:F11 User Guide]] and [[Category:F10 User Guide]] and F8 and F12 and ...
-Susan
-
Eric Christensen -
Ignacio Vazquez-Abrams -
John J. McDonough -
Karsten Wade -
Kirk -
Matthew Daniels -
Susan Lauber -
Wolf DreamWalker