Closed Bug 30653 Opened 25 years ago Closed 24 years ago

Search needs cleanup

Categories

(SeaMonkey :: Search, defect, P3)

Product:
SeaMonkey
Component:
Search
Type:
defect

Tracking

(Not tracked)

Status:
VERIFIED FIXED

People

(Reporter: braden, Assigned: blizzard)

References

Details

Attachments

(1 file)

This is a patch that gets rid of all the commercial crap in the URL history drop menu. Enjoy :)
16.09 KB, patch
Details | Diff | Splinter Review 
mozilla.org doesn't need to be pimping particular commercial search services--I
imagine commercial vendors are compensated to do exactly that. See bug 20862 for
precedent.
Blocks: 14532
moving to search
Assignee: cbegle → matt
Component: Browser-General → Search
QA Contact: asadotzler → claudius
So what are we doing about this search stuff defaulting to netscape?
People don't like it so much
One suggestion has been to default mozilla to dmoz.

A point to make clear:  the URLs defined inside of the included Sherlock files 
are used to determine when a HTML page is actually a "search result" to be used 
for the quick list in the search sidebar panel.  No Sherlock file for a search 
engine, no search sidebar results.

So, while the "mozilla.org doesn't need to be pimping search engines" is perhaps 
correct, its also worth while to note that by not having Sherlock files for as 
many search engines as possible, Mozilla users lose functionality.
Braden, please make a list of search services that you would be happy with, so
we can talk about some concrete suggestions.

The problem, as I see it, is the precedent being set by including commercial
services here. These are advertisements. Yes, they offer useful value. And from
the Idealistic Optimistic Capitalist point-of-view, *all* advertisements offer
useful value. Consumers, of course, would take issue with such a contention if
it were applied to real-world advertisements. But the bottom line is that the
utility of a particular advertisement is *always* a subjective call.

So where should the line be drawn? Whose advertisement gets to go into Mozilla,
whose is turned down, and on what grounds? The subjectivity involved makes this
is a slippery slope. Too slippery, IMO.

For this reason, I think it is a Bad Idea for Mozilla to include advertisements
for commercial services. And to that end, I think only other open source
products should get placement in the Mozilla UI.

To answer your question, Mike, the only service I know of that fits this
criterion is dmoz. I wish I could list some more, but I don't know of any.

Certainly having just one search service in Mozilla is an undesirable
limitation. Nonetheless, I'm of the opinion that the presently-employed
alternative sets too dangerous a precedent. The limitation to one default search
service could be substantially alleviated by allowing the user to add search
services. I'm not sure how practical that is, though.
OS: Linux → All
Hardware: PC → All
Personally, I don't know that I agree with the "search engine as advertisement" 
trivialization.

Its a disservice to Mozilla users to place such a harse limitation on the number 
of search services.  I believe a much more preferrable solution would be to 
encourage search services to play a part.  Otherwise, the search feature is 
irrelevant.

As another example, Alis has its language translation feature in Mozilla. They 
show branding when you try and translate a page from one language to another. 
Would you prefer to remove that useful feature as well?

I'd prefer for us all to "keep our eyes on the prize", which in this case I 
believe is enabling Mozilla to be as robust as possible in regards to internet 
search functionality.  (Naturally, along with keep legal issues in mind.)
"Trivialization"? If that remark is accurate, then it seems like it should be
"trivial" to produce a set of guidelines by which product placements could be
deemed appropriate or inappropriate for Mozilla. Such guidelines would most
likely put this issue to rest.

But please, don't ask me for help coming up with them. *I* don't think this
distinction is trivial.

And, yes, the Alis stuff is no less inappropriate. Can *anyone* submit a patch
to Mozilla inserting a link to their site in the browser's menus? If not, would
mozilla.org be able to objectively defend a decision to decline such a patch?
> Can *anyone* submit a patch to Mozilla inserting a link to their site in the 
browser's menus?

Anyone can *submit* a patch for anything.  Would the patch be accepted?  Its up 
to the module owner to decide the value of the submission.

Would I personally commit a patch which adds Joe-User's web page as a link into 
Mozilla?  Most likely not, assuming that it has little value to the average web 
user.  Would I commit a patch which adds a feature such as language translation 
or enhanced search support which thousands if not millions of people would use 
daily and find value in?  Quite possibly, yes.

You are entitled to your opinion.  As the module owner of search functionality 
in Mozilla, I have mine.
Indeed. And as an employee of AOL, you also have a conflict of interest.

These links represent the interests of your employer and those of entities with
whom your employer presumably has business relationships. It is certain that
having these links distributed in the Mozilla source as well as the Netscape
client presents additional benefit for AOL and additional value for these
business partners. Whether or not this situation impacts your decision on this
matter is irrelevant: a conflict of interest exists; there is the appearance of
impropriety.

If eliminating product placement in Mozilla is out of the question, the conflict
of interest could probably be mitigated with a mozilla.org policy outlining when
and where product placements are appropriate.

I would suspect that this (barely) subtle branding might make potential
competitors to AOL (or its affiliates) think twice before leveraging the Mozilla
code. That situation might be good for AOL. I can't imagine that it would be
good for mozilla.org.

<sigh>

Well, if product placement is indeed the direction Mozilla is going, I guess
this bug can be closed.

Its true that I am an AOL employee.

While you might see a "conflict of interest", you are missing the simple point 
that if you remove all the references to third-party search engines, then 
Mozilla's search sidebar panel becomes basically useless (unless users 
create/find search datasets and add them in by hand.)

I personally don't want to see that.

Why?

It means the feature is used less by developers, receives less quality, and the 
Mozilla browser is less usable.

I don't care if ANY AOL-sanctioned search engines are in Mozilla or not.

I do care that as many popular third party engines are made available (within 
the limits of the law) as possible.

Mozilla.org staff are working on this issue from a legal standpoint.

I am not missing that point. I just think it is decidedly secondary to avoiding
the potential for Mozilla to be reduced to a delivery mechanism for its
contributors' ads. But perhaps my views are alarmist.

But let me try to discuss a potential solution...

I am curious about the potential to enable users to add search engines
themselves, *almost* as arbitrarily as bookmarks can be added.

Looking at the Sherlock files, they seem pretty simple. I am inclined to suspect
we could come up with a pretty simple spec (relying on META tags and HTML
comments) by which search providers could create pages which a generated
Sherlock file could make sense of.

IMO, this opens up the process to the advantage of both users and search
providers. Users nolonger would need to be burdened with search services they
don't use, and search providers could make the change to enable this Mozilla
feature *entirely* on their end, so they wouldn't have to wait for Mozilla
source changes to trickle into the user base.

Does this sound like something that might be workable? If it does, I'll delve
further into this and work on such a spec.

See "http://www.mozilla.org/projects/search/technical.html" for an example of 
using JavaScript to add a search engine.
Note: "javascript:" URLs currently are broken (see bug # 34217).  If you want to 
work around it, you can use "onclick" handlers instead.
Unless I'm waay off base, isn't that *exactly* how it works now? The only 
question or discussion is which search engines would ship with the browser. 
Keeping in mind having one or none makes the feature less discoverable and 
providing the sherlock file w/o permission from the site is illegal.
Ah, cool! I didn't realize this much had been implemented. But no, what I
described is not "exactly" as it appears to work now, though there is obviously
substantial similarity.

Functionally, the thing that seems to be missing from the current implementation
is discoverability. I'm drifting beyond the scope of this bug here, but by
"discoverable" I mean that the browser could "look" at the page and see that it
would be a candidate for adding as a search engine. This might be useful to do
things like enable an "add" button in the search sidebar. Consider a combination
of META and LINK elements:

  <meta name="mozilla-search-name" content="Foo">
  <meta name="mozilla-search-category" content="Web">
  <link rel="sherlock" href="http://www.foosearch.com/foo.src">
  <link rel="sherlock-icon" href="http://www.foosearch.com/foo.png">

Perhaps something like that would make a good enhancement; but as has been said,
that's not the point of this bug.

IMO, the ability to add (and presumably delete) search services means it makes
less sense for Mozilla to proceed down what I perceive to be a slippery slope.
But I've already belabored that point and I won't do it further.

Since this bug hasn't been closed yet, I'm guessing that's because folks think
there's something amidst all this that ought to be addressed. Let's figure out
what needs to be done to move this bug toward closure. Removing non-open search
services appears to be out of the question. Matt wrote (about a month ago): "So
what are we doing about this search stuff defaulting to netscape? People don't
like it so much." The suggestion to default to dmoz sounds nice and
uncontroversial. Since mozilla.org apparently deems more drastic recourse
unnecessary, I'm of the opinion that that ought to close this bug.

I've considered META tags; however, JavaScript is the path that I've chosen to 
support due to various constraints.
I thought that I would insert some of my own opinion into this conversation.  I
also think that mozilla should not contain links to netscape services.  Example:
Clicking on the Search the Web on the Search menu goes to the netscape site.  It
should probably goto dmoz.org at the present time since that is the only
non-commercial search that I know of.

As far as the sidebar search panel goes, we should have links to most of the
major search engines.  They should not go through netscape in any way.  They
should be direct links.  The problem is determining the major search engines.
For now we should just pick some.  This maybe a good poll for mozillazine: What
are your favorite search engines, and then we use the top 10 or something.

Here are the ones on the bottom of search pages at dmoz.org.
All the Web - AltaVista - Deja - Google - HotBot - Infoseek - Netscape -
Northern Light - Yahoo

*** Bug 37843 has been marked as a duplicate of this bug. ***
Copying the dup bug 37843:

-------------------------- Reporter sford3@swbell.net ------------------------
It seems as if netscape shop has been added to mozilla, shouldn't this be in the 
commercial tree?



------- Additional Comments From leaf@mozilla.org 2000-05-01 18:46 -------

rjc, this stuff belongs in the ns tree somewhere.




------- Additional Comments From rjc@netscape.com 2000-05-01 19:12 -------

For everyone who DOESN'T want some search engines in Mozilla, there is someone 
else who does.  For example, see bug # 32586 for requests to add more engines.

I personally believe that Mozilla should accept search files from any 
official contributor (whether Netscape or anyone else) barring legal issues.  
So, if JoeUser submitted a search file for "Google" we should not accept it.  
However, if someone directly from Google submitted a search file for their web 
site, Mozilla should accept it.

Basically, the more search engines that are there, the more valuable the search 
sidebar panel is for the user.  If Mozilla basically has no search engines in 
it, then the search functionality in Mozilla won't function and will receive 
little testing.

Reassigning this bug to marketing, and cc'ing mitchell@mozilla.org



------- Additional Comments From mozilla@bucksch.org 2000-05-02 00:16 -------

Mozilla should propbaly include the top <= 10 search engines or so. (We propably
wouldn't want to have a Microsoft search engine (or something similar) in there,
even if it were part of the top 10 :) .)

Shop@Netscape certainly is not part of them. If you include Shop@Netscape, why
not ebay? Its competitors? Maybe amazon? egghead?

Please remove Shop@Netscape.

> [...] legal issues. So, if JoeUser submitted a search file for "Google" we
> should not accept it. However, if someone directly from Google submitted a
> search file for their web site, Mozilla should accept it.

What legal issues? It is OK to include a search engine in the bookmarks by
default. HTML clients have the freedom of presentation. => JoeContributor should
should be able to submit new search files. If they are accepted depends on the
Mozilla community.

-- Additional Comments From amitp@google.com 2000-04-27 12:59 in bug 32586 --
> One problem seems to be that the search engine companies aren't going to
> guarantee that their HTML is formatted in any particular way.  If Netscape
> distributes Sherlock plug-ins for search engine XYZ, and XYZ changes its
> HTML, then users will complain to Netscape.
> That's no reason not to put them in Mozilla, though.  :-)



------- Additional Comments From rjc@netscape.com 2000-05-02 01:33 -------

> What legal issues?

One example: Many major search engines strictly prohibit redisplaying of their 
copyrighted content in any form other than what they present it as.



------- Additional Comments From mozilla@bucksch.org 2000-05-02 01:59 -------

> One example: Many major search engines strictly prohibit redisplaying of their
> copyrighted content in any form other than what they present it as.

How do "they present it as"? There's no "correct" way to display HTML. It is by
definition completely up to the client, what to do with it.



------- Additional Comments From rjc@netscape.com 2000-05-02 02:37 -------

I'm not going to argue this point. Its merely a fact. We do re-display search 
results in an altered form.



------- Additional Comments From braden@endoframe.com 2000-05-02 10:57 -------

Folks, most of this discussion has already taken place in bug 30653. (Is this a
dup?)

BenB: Top ten? That's a totally arbitrary distinction. On what logical,
defensible basis would you rationalize not including a particular search
service? And why *wouldn't* we want a Microsoft search engine??? It appears to
be okay with you for mozilla.org to have overt prejudice about who will and
won't go into the Search feature. While I'm not happy with the situation as it
is now, we *definitely* don't want to go in that direction!

-------------------------------------------------------------------------------
My new comments:

braden, yes, it is arbitarily, but (nearly) the same as David suggested. Dunno,
if it's worth a poll. Maybe.
Please note, that dmoz' search engines at the bottom are propably also
influended by Netscape: it included many more 2 month or so ago, including my
favorite one: MetaCrawler.

rfc wrote at 2000-04-09 04:19:
> Would the patch be accepted?  Its up 
> to the module owner to decide the value of the submission.

Please keep in mind, that the Mozilla community has the final call.
Jeremy, this would be bug 37873.


Jeremy, it looks to me, as if it were not a patch, but a full source file.

Please attach a diff (|cvs diff -u <sourcefile> > <difffile>|) to bug 37873. Tnx.


Don Melton says, "What should we do with this one/who should own it?"
Assignee: matt → rjc
.
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → LATER
I don't think, this should be resolved later. The thread in .seamonkey shows,
that this needs to be resolved, and I am not the only one thinking this.
REOPENing.

We need a strategy, how to decide, which ones are to be included and what we
will do, if a search engine provider complains. This should be decided by
mozilla.org/the mozilla community.
Status: RESOLVED → REOPENED
Resolution: LATER → ---
REASSIN to shaver@mozilla.org
Assignee: rjc → shaver
Status: REOPENED → NEW
There are two basic opinions:
- It's only there for raw testing, so any will do.(practically: AOL's services)
- For Mozilla dogfood, we need to have a reasonable default set, where
reasonable means acceptable for the Mozilla community
  - No commercial engines at all
  - A limited set of popular ones
  - All

My opinion: I think, the last one is impossible. 1. ignores dogfood and also
looks bad for an open source project.
> the last one

the very last one, i.e. "All"
La la la, pass the buck.
Assignee: shaver → blizzard
For the sake of testing ( and because we have permission ) I've checked in a
google search source.
Status: NEW → RESOLVED
Closed: 24 years ago → 24 years ago
Resolution: --- → FIXED
marking VERIFIED
Status: RESOLVED → VERIFIED
*** Bug 37315 has been marked as a duplicate of this bug. ***
Product: Core → SeaMonkey
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: