Open Bug 757736 Opened 13 years ago Updated 4 months ago

[Meta] Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options (e.g. tagging system)

Categories

(Thunderbird :: Address Book, defect)

defect

Tracking

(Not tracked)

People

(Reporter: it, Unassigned)

References

(Depends on 8 open bugs)

Details

(Keywords: dataloss, meta)

User Story

Blocking Bug 75711 describes one promising approach/solution for fixing the conceptual problems of the current mailing list design as described and discussed in this bug 757736, together with (various) possible solutions.
User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:12.0) Gecko/20100101 Firefox/12.0 Build ID: 20120420145725 Steps to reproduce: Mailing lists are certainly not one of the most prominent feature of TB, but everyone how tries to use them seriously faces major issues, such as: Actual results: 1. When viewing or editing an address book entry, one cannot see/edit in with lists(s) the entry is a member. This makes list membership checking very awkward and is a pitfall, e.g., list membership info may be lost by deleting the wrong duplicate entry (see bug 756840 and bug 654744). 2. When deleting an address book entry, all mailing lists referencing it get a "hole" at the place where its primary email address was shown. 3. After changing the primary email address of an address book entry, double-clicking a list name containing the entry, and clicking "Ok", a spurious address book entry appears with the old email address. For sure also other further bugs have been reported that are due to the bad design. General comments and analysis: Whoever introduced mailing lists in TB apparently did not think more than five minutes about this concept. Their design is painfully naive, namely a list of email addresses, where the address book entries are associated/referenced via their primary email address field. Developer documentation on the mailing list concepts and their implementation is very hard to find (where is the definite source?) and all what I actually found is extremely poor, outdated and confusing. Moreover, its implementation is pretty awkward and still very buggy. Expected results: The list membership association *must not* be done via any user-defined field, but (if at all) via the UID (BTW, does it currently exist as an implicit/automatic field?) of the address book entry. This would solve issue 3 given above. I strongly advise re-designing mailing lists, reversing their reference concept: namely to allow storing (or referencing) in each address book entry multiple user-definable tags. In its simplest use, each such tag is a replacement for a mailing list name and can be used to directly address a group of receivers bearing the given tag. This straightforward and popular mechanism already cleanly solves all the three issues mentioned above. A very nice more general use of this tag concept is that it would allow for a very flexible and convenient on-the-fly creation of mailing lists via logical expressions, e.g.: "send mail to all entries that have both tag1 and tag2, but not tag3". Such an expression could be placed in each To/Cc/Bcc field, where the effect of an outermost logical "or" operator can then be achieved by using more than one To/Cc/Bcc field.
Adapted platform and severity tags, added a keyword and a few bug references.
Severity: normal → major
Keywords: dataloss
OS: Mac OS X → All
Hardware: x86 → All
See Also: → 654744, 756840
See Also: → 353563, 173836
For meta bugs, it's better to add bug references in depends field, because: - see also links are not simplified yet - see also links don't show bug summary and status yet - see also links do not create a reciprocal link on the target bug yet (For all of these problems, bugs have already been filed) In contrast, dependencies are simplified, show bug summary & status, and create reciprocal link on target bug.
Status: UNCONFIRMED → NEW
Depends on: 173836, 353563, 654744, 756840
Ever confirmed: true
Keywords: meta
See Also: 173836, 353563, 654744, 756840
(In reply to Thomas D. from comment #2) > In contrast, dependencies are simplified, show bug summary & status, and > create reciprocal link on target bug. ...and can be comfortably viewed in dependency tree/graph (see links at top of bug)
David, I understand that this meta bug intends to collect the following subset of mailing list related bugs (pls correct me if I'm wrong): Bugs that are about the *conceptual design* of mailing lists, limited to those regarding the *relationship between cards and mailing lists* - shortcomings of the current conceptual design of that relationship - suggestions for improvement I recommend to understand this as a narrow definition: - avoid adding *all* of the mailing list related bugs here, currently around 100: These are better found with dynamic search from comment 4 (because all mailing list bugs should have "mailing list" in their summary). - avoid adding technical bugs here, unless they are directly related to conceptually bad/wrong design of relationship between cards and mailing lists
I'm a bit concerned that this meta bug mixes general problem with specific solution proposed in comment 0, and I see some inconsistencies. But of course, you are free to define meta bug with your own ideas and intentions. Generally speaking, I would avoid specific solutions, especially in comment 0 of meta bug. It's better to add them in subsequent comments or even in separate bugs (which imo not necessarily need the meta keyword if the summary transports the more general intention, but I might be wrong). Perhaps we can rephrase the summary to narrow down the problem while opening up for other potential solutions: Suggested summary for this meta bug: "Mailing lists' conceptual relationship with cards is fundamentally flawed - explore options of improvement" If possible, we should also decide the general intention/direction of this meta bug: Is it... - tracker bug that mainly collects problems and solutions of dependent bugs (without much discussion here) - intended for discussing problems and developing solutions right here - or both?
Thanks Thomas for your structural improvements and various comments. (In reply to Thomas D. from comment #5) > David, I understand that this meta bug intends to collect the following > subset of mailing list related bugs (pls correct me if I'm wrong): > > Bugs that are about the *conceptual design* of mailing lists, limited to > those regarding the *relationship between cards and mailing lists* > - shortcomings of the current conceptual design of that relationship > - suggestions for improvement Yes. (In reply to Thomas D. from comment #6) > Generally speaking, I would avoid specific solutions, especially in comment 0 of meta bug. > It's better to add them in subsequent comments or even in separate bugs I agree that specific, detailed solutions should not be the topic of this meta level bug, but here we should concentrate on the fundamental problems and conceptual solution options. > Perhaps we can rephrase the summary to narrow down the problem while opening up for other potential solutions Good suggestion - I re-phrased the summary along the lines you proposed.
Summary: [meta] mailing lists are fundamentally flawed - better use tagging mechanism → Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options
(In reply to Thomas D. from comment #6) > - tracker bug that mainly collects problems and solutions of dependent bugs > (without much discussion here) Difficult to fulfil, if I don't have right to add a bug to the depends on list ;-) Another candidate: bug 173798
Blocks: 757900
Another candidates: bug 189895, bug 628035
Depends on: 757931
(In reply to Ulf Zibis from comment #8) > (In reply to Thomas D. from comment #6) > > - tracker bug that mainly collects problems and solutions of dependent bugs > Another candidate: bug 173798 (In reply to Ulf Zibis from comment #9) > Another candidates: bug 189895, bug 628035 Bug 173798 is currently expired, so there's not much point in adding that. Even if it were reopened, I wouldn't add it here because imo it is not related to the wrong design of relationship between mailing lists and cards. As explained in comment 5, let's avoid to add technical bugs that could be solved without changing the current concept of mailing lists, because the list of technical bugs in mailing lists is too long to add them all, and there's no point because they can reliably be found with simple search (see link in comment 4). Same for Bug 628035: technical problem (missing screen update of contacts list pane) which is independent of the conceptual problem. The conceptual problem is something like Bug 85687 (why create extra cards at all?) - I'll add that here - and Bug 654744 (in current concept, need to warn users that deleting cards affects mailing lists) - already dependent here. Bug 189895 is the type of bugs which we want to collect here, so I'll add that, too. I'm a bit strict on these because a limited, focused choice of dependants will be best to make this meta bug more useful.
Depends on: 85687, 189895
Depends on: 64169
Depends on: 75711
For search results, we also need [Meta] in the summary.
Summary: Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options → [Meta] Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options
I filed the following Meta bug with a smaller, more technical, less conceptual scope (intersections with this meta bug might occur), because in spite of conceptually wrong design (this bug 757736), a lot of current problems would disappear if following problem would be fixed: Bug 758969 (1) - [Meta] Mailing Lists: Bugs caused by problem: Contacts list pane of contained contacts not updated/refreshed (artefacts, duplicate entries, apparent and real dataloss of cards, display names etc.) (1) https://bugzilla.mozilla.org/show_bug.cgi?id=758969
This is a noble effort, but I'd hate to see you and others potentially wasting your time here, given (I think / my impression is) many mailing list issues are overly challenging and unlikely to be fixed until the AB redesign delivers entirely new AB backend. Or a trivial patch likely / a volunteer likely to step forward soon to fix?
Depends on: 118270
Depends on: 136757
Summary: [Meta] Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options → [Meta] Relationship of mailing lists with address book cards is conceptually flawed - explore improvement options (e.g. tagging system)
See Also: → 1193290
Depends on: 1329295

I think this is a duplicate of bug 75711.

See Also: → tb-ab-rewrite
See Also: → 1777474
Severity: major → S3

(In reply to Magnus Melin [:mkmelin] from comment #14)

I think this is a duplicate of bug 75711.

Closely related, but not a duplicate imo.

  • This bug 757736 is a general meta bug which describes the broken relationship between mailing lists and cards (iow, our current deficient way of creating mailing lists), and seeks to explore possible solutions (open-minded).
  • Bug 75711 looks like one particular promising approach/solution for this bug, but until that is fully understood, explored and declared as the solution which we really want to fix the mailing list problem, I think this bug 757736 should remain separate.
User Story: (updated)

Well, what is the competing alternative solution?

You need to log in before you can comment on or make changes to this bug.