Just a few of the bugs relating to problems with text in Firefox:
https://bugzilla.mozilla.org/show_bug.cgi?id=79864
https://bugzilla.mozilla.org/show_bug.cgi?id=180625
https://bugzilla.mozilla.org/show_bug.cgi?id=470384
https://bugzilla.mozilla.org/show_bug.cgi?id=470642

The 2001 bug in Mozilla core is probably a blocker to problems with Diigo such as
http://groups.diigo.com/Diigo_HQ/forum/topic/9245

These bugs and others in Mozilla are IMHO strong reasons to:

1) _not_ accept, as a response, "give Firefox a try"†

2) encourage Diigo services to be as far as possible web compatible, browser-agnostic.

Best
Graham

† http://www.diigo.com/04jwx

Subject: enter and tab keys for completion of tags

When some things are presented as menu items, it's proper for the enter key to select the highlighted menu item.

If the right arrow is pressed whilst a sub-menu is selected, the selection moves into the sub-menu.

From that point, the left arrow moves from the sub-menu to the parent menu.

Tabbing moves from one main menu to another, so for example I can select anything in a File menu; then the tab key will move the cursor right and up, to the head of the Edit menu; and from any point in the Edit menu, shift-tab will move left and up to the head of the File menu.

Key navigation of menus should be no different in a Firefox menu, no different in a Diigo menu.

See also http://groups.diigo.com/Diigo_HQ/forum/topic/acceptance-of-a-tag-suggestion-is-misinterpreted-too-soon-as-a-new-topic-submission-7165

Regards
Graham

KGyST1 KGyST wrote:

> I'm afraid that You misunderstood Me.
>
> I don't speak about menus in anywhere

I did misunderstand, apologies, thanks for clarifying!

>> an autocomplete list rolls down with the suggested tags

Elsewhere in Diigo, some places:

* a list of suggested tags rolling down _is_ a menu

We realise now that in some areas, including (maybe not limited to) Diigo 3.1.6.13 bookmark dialogue in Firefox:

* a list of suggested tags rolling down is _not_ a menu.

Presentation is very different in the Diigo dialogues with which I'm familiar, hence some confusion ;)

KGyST1 KGyST wrote:

> I don't want to order anything by frequency, but by *Frecency*.

OK, now I see http://en.wiktionary.org/wiki/frecency and get it

(To be honest, I read 'KGyST1 KGyST' and thought 'frecency' was a typo!)

> Thanks and Happy new year,
> Sam

And to you, and to other readers - cheers - Graham

Thanks to participation in another list, I discovered Fitts's law.

Reading no more than the first two paragraphs of http://en.wikipedia.org/wiki/Fitts_law I guess that this law, or something like it, should (if possible) be applied to issues highlighted in two topics:

http://www.diigo.com/annotated/0a14f1af9178ca9f011ad17a287c4561
http://www.diigo.com/annotated/e8c0911164affe88fa14c39f8d428909

The formulas associated with the law scare me (!) but AFAICT the gist of it is, to have the gestures (mouse/cursor movements or whatever *feel good*.

Example: http://www.zotero.org/support/zotero_over_a_network



AFAIR this is the first site at which I have found (within the site) no hint of what's hidden whilst Diigolet is present.

(We see that the word 'ZOTERO' is cropped but there's no hint that Login, Register and Search are obscured.)

= Consideration =

Some pages are presented with Diigolet loaded automatically. Example:
http://www.diigo.com/04j3t.

= Distinction =

The New York Times situation http://groups.diigo.com/Diigo_HQ/forum/topic/new-bookmarklet-not-working-with-the-new-york-times-4368 differs; there, apparently it is a feature of the site that prevents use of Diigolet. Here, it is a feature of Diigo that obscures a part of a site.

= Focii of this topic =

Not the Zotero site in particular.

Focused on a list (hopefully very short) of sites at which the presence of Diigolet obscures something that is primary, or is of importance to the user.

Postscript: added an example.

@ Diigo only

Observe the URL discrepancy in this screen shot:



- that is, http://slides.diigo.com/list/grahamperrin/ … as a prefix during _edition of a list_.

With my oldest list:

* I tried twice to reproduce the issue, could not reproduce.

With my most recent list (created today):

* the issue is reproduced:

1. view a list that you created

2. accept a WebSlides view of your list

3. Play

4. collapse

5. You are here: Diigo Home > Webslides > …

6. click Webslides

= Issue =

slides.diigo.com domain in Diigo list context

= Scope =

I imagine an issue arising only if relative linking is used around this area.

The first few functions that I tried, from this 'odd' perspective to (URL for) a list were fine, so I do not treat this as a bug.

= Summary =

Reference item; to be considered if a user reports issues whilst editing or working with a Diigo list.

> 4. collapse
>
> 5. You are here: Diigo Home > Webslides > …
>
> 6. click Webslides

Sorry, a mistake whilst drafting. The step beyond (4) to reproduce was more like:

5. follow the link that is uppermost in the WebSlides window.

However: whilst the issue *was* reproduced earlier, I can't reproduce it now. I tried a few variations of step (3) (Play) and beyond, but all paths lead to a proper http://www.diigo.com/list/ … prefix to the URL for the list.

= Conclusion =

Random or unpredictable.


Postscript: correction.

Bug

http://slides.diigo.com/list/grahamperrin/webarchive presents my personal profile.

Steps to reproduce

1. visit the WebSlides URL

2. at the beginning or end of the show, click By Graham Perrin

Is it possible for WebSlides to present the results of a feed, a feed that is is served by Diigo?

I imagine something like, a Diigo list that updates itself:

* periodically

or

* at the request of the author of the list

- to comprise the contents/abstract of the feed.

For this type of list, I would understand that associated Diigo highlights and annotations change over time (whenever the URLs associated with the list change in harmony with the feed).

I should not expect the list to grow infinitely. I imagine:

* a limit set by Diigo on the number of URLs that this type of list may comprise

and

* a user preference to have a limit lower than the one set by Diigo.

Limits: to discourage misuse and respect Diigo service load, performance, terms etc..

> I imagine something like, a Diigo list that updates itself:

> or
>
> * at the request of the author of the list

If at the request of the author, on demand (not periodic, automated) then the routine could be a sibling to
http://www.diigo.com/tools/import_all

I imagine importing a feed to a list, with the option of overwriting the contents of an existing list.

Hi Onno

Your provided tag allows you to find other topics similarly tagged:

http://groups.diigo.com/Diigo_HQ/forum/tag/api

The URL offered requires authentication so I can't view its content.

Does Diigolet respond with a specific message?

If I attempt to use Diigolet at
http://www.extranetevolution.com/extranet_evolution/2007/03/wikis_and_aec_c.html
then Diigo responds:

> :( Diigolet doesn't support this page.

http://groups.diigo.com/Diigo_HQ/forum/topic/8988#4 made me wonder whether any users wish for multiple personas.

If I gained multiple personas, then I might have amongst my dashboard items

* few most recent bookmarks by http://www.diigo.com/user/fuzzy
* few most recent bookmarks by http://www.diigo.com/user/grahamperrin

(AFAIR fuzzy was my previous pseudoynm for which I foolishly forgot the password. That's OT; now in reality I'm happy with the one name.)

Put another way: I might

1. choose a persona at time of bookmarking/highlighting/annotation (select one of two personas from a pull-down menu, or click one of two radio buttons alongside personas)

2. occasionally seek an overview/sense of what has been bookmarked by both personas

Personally, I might choose to offer a public overview of both personas.

Marsh Feldman might have personas 'Teacher Feldman' and 'GeorgeW Feldman' and Marsh Felman would probably enjoy an overview of both personas, whilst never allowling a student to become aware of that overview or the association between personas.

The words "relational databases nightmare" come to mind ;)

I do have an approach to personas that need not bring database administrators to their knees. I should aim to share that approach in a separate topic.

Maggie Tsai wrote:

> we may be able to release a feature to allow users to change
> username soon

Please, will Diigo 4 allow this?

Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8.1.19) Gecko/20081204 SeaMonkey/1.1.14

1. http://www.diigo.com/tools

2. the Mozilla browser is recognised and so Diigo Toolbar 3.1.6.13 is offered for installation

= Bug =

> firefox diigo toolbar
> install script not found

3. at https://addons.mozilla.org/en-US/seamonkey/ for SeaMonkey

4. seek and find Diigo https://addons.mozilla.org/en-US/seamonkey/search?q=diigo&cat=all

5. Add to Seamonkey

= Bug =

> Sorry, you need a Mozilla-based browser…

(Seamonkey _is_ a Mozilla-based browser)

7. click OK

8. click Install

= Bug =

> diigo-20081015.xml
> Not a valid install package

9. http://www.diigo.com/tools?client=flock | Install Diigo Toolbar 3.1.6.13

(Flock is another Mozilla-based browser)

= Bug =

> flock diigo toolbar
> install script not found

= Workaround =

* ignore invitations to add/install Diigo Toolbar

* use Diigolet 3.1b446 or greater.

Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.2a1pre) Gecko/20081220 Minefield/3.2a1pre

1. http://groups.diigo.com/Diigo_HQ/bookmark

2. tick | More Actions… | Extract Annotations…

3. click the Copy button

= Bug =

The copy does not occur.

4. http://www.diigo.com/ | My Bookmarks

5. Share menu | Get Annotated Link…

6. click the Copy button

= Bug =

The copy does not occur.

7. Share menu | Extract Annotations…

8. click the Copy button

= Bug =

The copy does not occur.

@ Joel

Agreed, good idea.

Considering
https://bugzilla.mozilla.org/show_bug.cgi?id=135268#c79 reference to
https://bugzilla.mozilla.org/show_bug.cgi?id=470642 I wonder whether the symptom is somehow a side effect of novel handling of strings.

Re http://www.mozilla.org/developer/#builds I might test a nightly for Linux (on Ubuntu 8.10) or Windows (in Darwine).

testing www.diigo.com with Firefox nightly

9. click the 'Comment' button within Diigolet toolbar

10. Add

11. in the comment, enter some text with a hyperlink, for example

testing www.diigo.com with Firefox nightly

12. Post

13. attempt to select the text within the note

14. attempt to follow the link within the note

= Bugs (apparently the same for comments as for sticky notes) =

Text can be neither selected nor copied.

Link can be neither followed nor copied.

= Observations =

Within the sticky notes window, and
within the comment window pane to page comments:

a) service-defined links succeed

b) user-authored links fail

Within the comment window pane to sticky notes:

c) service-defined links succeed

d) user-authored links are ignored; a
service-defined link presents the related sticky note (logical :)
with user-authored links (currently unusable).


= Suggestions =

1. Review http://groups.diigo.com/Diigo_HQ/forum/topic/comments-hyper-links-missing-angle-brackets-misinterpreted-line-breaks-missing-9015 and related topics

2. Reconsider the re-introduction of reasonable (not excessive) rich text edition within sticky notes and comments

3. Respect and understand the good behaviour of point (d) above.

Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.2a1pre) Gecko/20081220 Minefield/3.2a1pre

Result:

Diigo Toolbar 3.1.6.13: not compatible.

The majority of other add-ons: not compatible with Firefox.

@ Joel: thanks, resolved.

I'll make no further attempt to test Diigo Toolbar with any version of Firefox 3.2 until after http://blog.diigo.com/category/toolbar-change-log/ suggests compatibility.

GUIDELINES

This topic in the Diigo Community forum is a place for
open discussion that relates to
ALL THREE of the following:

1. Mozilla Firefox
2. Apple System Services on Mac OS X
3. Diigo.

BUG REPORTS

This topic is intended for general discussion of the three points above, in combination.

If you have any bug report, please express it in a new topic. Thanks :)

RECOMMENDED READING

http://www.brighton.ac.uk/centrim/Members/gjp4/2008/12/2008-12-21 is my blog post,

> Mozilla Firefox and System Services: state of the union

RELATED

Nearby http://groups.diigo.com/Diigo_HQ/forum/topic/9252 is my plea for Bugzilla@Mozilla users/members to add votes to relevant bugs.

If you have an account with Bugzilla@Mozilla and a free vote (I had 1,000 votes to cast when I joined) then please, can I encourage you to cast one vote for one or more of these three bugs:

https://bugzilla.mozilla.org/show_bug.cgi?id=104331
https://bugzilla.mozilla.org/show_bug.cgi?id=367867
https://bugzilla.mozilla.org/show_bug.cgi?id=394599

Bug 104331 was reported in 2001

/cough

and the absence of Firefox support for System Services (interapplication communication) is not an encouragement to developers.

My own votes have been cast :)

For those of you who prefer a Gecko engine to a WebKit engine: Gecko-based Camino has been around since February 2002. One of the driving forces behind Chimera (as it was originally called) was Firefox lack of support for System Services.

It's seven years since the bug was reported to Mozilla developers, and nearly six years since Chimera/Camino became the faster, better, System Services-enabled citizen on Mac OS X, so Firefox is _really_ behind the times on this one.

On the bright side: I see a small flurry of December 2008 discussion in the 2001 ticket, so maybe the Mozilla developers are close to a first breakthrough :)

Many thanks
Graham

https://bugzilla.mozilla.org/show_bug.cgi?id=470642
is my request for enhancement relating to loss of hyperlinks and other richness.

https://bugzilla.mozilla.org/show_bug.cgi?id=470651
is my request for enhancement relating to loss of images.

Keywords for both bugs:

interop

As far as I can tell these losses, whilst clearly undesirable, are _expected_;
http://hg.mozilla.org/mozilla-central/rev/923d927753ce#l1.42
acknowledges a limitation 'sending strings to services' and
https://bugzilla.mozilla.org/show_bug.cgi?id=135268#c64
acknowledges that the resolution excludes images.


Postscript: beginning parallel to this thread, there's a separate topic
> Mozilla Firefox, System Services and Diigo: state of the union and focused discussion

Mozilla bug 470642 seems to be not far from resolution :)

If you're a user of Firefox, Flock etc. on Mac OS X, then it's well worth voting for Mozilla core
bug 79864 - Mac style text 'styl' or RTF support for mozilla clipboard

> Assigned To: Nobody; OK to take it and work on it

- if you know any developer with relevant skills, you might like to steer them in that direction :)

- for as long as bug 79864 remains open, it's impossible to copy rich text or HTML from Mozilla applications on this platform (AFAICT the clipboard in Mozilla is limited to plain text).

---

Another one that's worth progressing, or at least voting for is Mozilla core
bug 470651 - System Services (interapplication communication on Mac OS X): services to support images in Mozilla applications

- I wonder whether that will be blocked by bug 79864.

1. http://www.diigo.com/ | My Bookmarks

2. add a private comment to the bookmark

3. add a public comment to the bookmark

4. Share menu | Extract Annotations…

= Expected =

Within the extraction: the public comment.

= Bug =

Extraction has no annotations.

= Screen shot =



= Environment =

* Safari 3.2.1 (5525.27.1)
* Mac OS X 10.5.6, Intel
* ISP speed test results http://www.thinkbroadband.com/speedtest/results/id/122979590721925427608.html immediately following the bug look OK, probably irrelevant in this context

Seven minutes later: a reload of the page and a repeat of step (4) but still no public comment in the extraction:



It certainly worked in the past, and at a glance it seems to extract what's intended within the Diigo groups UI:



and at a glance the extraction is successful for most annotated public bookmarks of my own:



so maybe for the most recent annotations, some waiting time is required ;)

For anyone who is wondering whether

* the primary link presented within an extraction of annotations

is the same as

* one of the the four types of annotated link

the answer appears to be:

* yes, the same.

= Example, using Diigolet =

1. My Bookmarks | aim for a particular bookmark

2. Share menu | Get Annotated Link… | ignore the two check boxes (the other three options) | ignore the Copy button | select then copy the URL | close the dialogue

3. Share menu | Extract Annotations… | follow the primary link | copy the URL from the address bar of the browser

Result of both 2 and 3:
http://www.diigo.com/annotated/04546b8def0580f1e6adb9de48f07df3

1. read and annotate the first in an ordered series of pages, introduction page A

2. read and annotate the second in an ordered series of pages, main page B

3. read and annotate third in an ordered series of pages, summary page C

4. read and annotate third in an ordered series of pages, conclusion page D

5. http://www.diigo.com/ | My Bookmarks

6. select the three bookmarks that you read in an orderly fashion

7. Share menu | Extract Annotations…

= Expected =

Introduction, main, summary, conclusion.

= Bug =

Extraction is conclusion, summary, main, introduction - in the wrong order.

= Screen shot =



- that's a genuine example of what should be last (the [SUM]) appearing first.

I intended to use (not simply play with) those annotations, but as they were disorderly I composed by hand instead.

= Suggestion =

Allow manual ordering of bookmarks.

I'll pretty much rule out this suggestion; we have lists, which can be ordered.

= Workaround =

1. create a list

- no, I'll stop there, this has the potential to 'waste' URLs

- and in any case the workaround would be long winded and require housekeeping (deletion of the list) after the routine.

= Suggestion =

1. assume that reading is orderly

2. assume that the related bookmarking is orderly

3. extract annotations with a normal reading chronology (the end is last) based upon the order of the bookmarks.