News for hackers. Stuff that matters.
Re:Work on these things by wslh on Thursday January 09, 2020 @08:59AM (#22001188)
I would add new ways of searching in Internet. If I searched for food recipes in the 2000s I would find independent blogs with some real local/family taste. Now I have a hundreds of results from click bait sites with the same commoditized recipes and the ugly blog with a good recipe deep in the long tail. We can say that we need improvements in the long tail when the tail deserves to move up (or to the left in a xy chart).

Internet search is a driver for the world economy, a tiny improvement would improve the life of entrepreneurs and their ecosystem beyond elite circles.

Re:Work on these things by _fs on Thursday January 09, 2020 @09:17AM (#22001334)
Are you saying that you don’t enjoy the 4 page life retelling story intermixed with countless ads and the same staged photo from 5 different angles and 4 different focal points when you just want to see the recipe? Luckily, there is an extension to parse all that out.

https://github.com/sean-public/RecipeFilter

Re:Work on these things by zozbot234 on Thursday January 09, 2020 @09:12AM (#22001295)
The problem with this is that black-hat SEO's have learned to mimic "independent blogs" and their linking/referencing networks (blogrolls, etc.) to the point that search engines are basically forced to discount all of these results. We need trusted, curated guides to the "independent" Internet, like the ODP/DMOZ of old. For both humans and search engines alike.
Re:Work on these things by PaulHoule on Thursday January 09, 2020 @09:26AM (#22001414)
Google in particular doesn't have an incentive to direct people to the independent web. Every year, there are fewer organic links in search and more ads and more links to Google properties.

http://www.seobook.com/brands-vs-ads

One of the worst changes is their effort to hide Wikipedia results for anything health-related and instead send you to "reputable" sites that will hit you with multiple pop-ins, heavy advertising that harms the reputation of the sites, and content sliced thin like salami to increase page views and ad impressions.

The same is also true of "reputable" sites like anandtech, which post good articles, but have the problem that there are so many ads on the site that the ads interfere with the other ads on mobile.

So far as ODP/DMOZ, that ran into the problem that the editors became corrupt and pretty soon the only way to get into the directory was to bribe an editor.

It's a tough problem.

Re:Work on these things by pcmaffey on Thursday January 09, 2020 @10:31AM (#22001957)
Ive got a half baked project for a blog search engine that won’t index any pages with ads...
Re:Work on these things by yifanl on Thursday January 09, 2020 @02:56PM (#22004944)
How do you define "ad"?

Keep in mind that as soon as you do define it, people are going to look for a workaround.

Re:Work on these things by _0ffh on Thursday January 09, 2020 @03:24PM (#22005260)
I suppose banning everything that pulls in stuff from any of the big ad-pushing companies would be quite a good start. You could probably even use existing open source ad-blocker blacklists. The only workaround from the top of my head is self-hosting advertisements, and there's a reason why this is not the norm.
Re:Work on these things by pcmaffey on Thursday January 09, 2020 @04:43PM (#22006199)
Yes, using an ad-blocker list is exactly what I was planning. Would then blacklist from the index any sites that serve ads. Would prolly also at a minimum, heavily penalize affiliate links.
Re:Work on these things by _0ffh on Thursday January 09, 2020 @07:56PM (#22007546)
Sounds like a worthwhile project to me! Don't forget to show HN in case you go ahead with it! ;-)
Re:Work on these things by bostik on Thursday January 09, 2020 @03:17PM (#22005186)
> How do you define "ad"?

Attempt to sell.

Whether that's for an agenda, a product, a way of life, service, access, or anything else with a direct (or indirect) transaction of value - it's a dead giveaway.

Re:Work on these things by yifanl on Thursday January 09, 2020 @03:36PM (#22005403)
Is a recipe for my grandma's minestrone soup (which requires you to buy a specific brand of chicken stock) selling you something?

Is a blogpost highlighting a particularly efficient way to write a tight loop selling you something?

Is a plain white webpage selling you the benefits of that particular shade of white?

Re:Work on these things by bostik on Thursday January 09, 2020 @04:06PM (#22005758)
I'll take your questions at face value. You are likely being downvoted for your use of an aggressive tone, but that's another matter.

1) Recipe for a good minestrone soup, no. Actively endorsing a specific brand of an ingredient, yes.

2) No. (One might argue you're building a brand and advertising yourself. I don't, necessarily. If the efficient loop was possible to create using only a particular compiler, then yes.)

3) Yes. The page is selling a particular idea, and most likely lying by accident. After all, how many screens have been calibrated for colour correction accuracy? Are you, as a reader, getting the correct information?

Re:Work on these things by yifanl on Thursday January 09, 2020 @04:39PM (#22006134)
Apologies for the brash tone.

I take the view that making a value judgment on an idea is essentially selling that idea. i.e. "A is better than B" is clearly advocating for A, as is "A is good"

I disagree with ads in general, but my problem is that blocking any form of endorsement as you suggested above would mean we can't transmit really any useful information. Obviously that means there's some acceptable level of endorsement... but that goes back to the problem of having a game-able filter.

Re:Work on these things by bostik on Saturday January 11, 2020 @04:20AM (#22019039)
Don't worry, I am well aware that my answer came off as snarky and brash so a similarly phrased response should not have been unexpected. And I know that my position is extreme. Even with the overall attitude change over the past 6-8 years about advertising[ß] my views are still outside the mainstream.

This is a personal opinion, but I believe you can do endorsements without falling into advertising trap. It takes effort, because you need to quantify your reasons of choosing a particular option (or a set of options) over the remaining ones. You need to specify the reasons, be painfully aware of and outline your own biases, and be willing to acknowledge the tradeoffs in your stated choice.

Obviously, if you are externally incentivised to make a particular choice before you broadcast about it, that's an ad. As one may imagine, I absolutely detest the entire influencer phenomenon - to the point of treating any product or service using them as bad actors perpetuating a disgusting system.

ß: Back in 2011/2012 HN was, in general, very pro-advertising. A lot of the [vocal] posters made their money on it or dealt with large advertising budgets on a daily basis. Ad block users were derided as parasites or thieves. Anyone questioning the common good was pushed away. Compare to the tone shift over the past couple of years and the difference is stark.

Re:Work on these things by sincerely on Thursday January 09, 2020 @04:02PM (#22005707)
No, no, no
Re:Work on these things by PaulHoule on Thursday January 09, 2020 @04:32PM (#22006039)
Many of the worst mind-polluting ads aren't trying to sell anything explicitly. For instance ads that talk about what an innovative and young company GE is, or how ADM is Supermarket to the World, etc.

PSAs also often fit into the same category.

Re:Work on these things by willismichael on Thursday January 09, 2020 @03:45PM (#22005496)
Well, there will only be a major effort to find a workaround if it goes mainstream.
Re:Work on these things by bjornlouser on Thursday January 09, 2020 @11:00AM (#22002231)
"Jerry's guide to the world wide web"
Re:Work on these things by gombosg on Thursday January 09, 2020 @09:11AM (#22001286)
IMHO this is because of SEO. Whatever algorithms search companies make, those who exploit them (called SEO, whatever the hat color is) will have an advantage over those who don't.
Re:Work on these things by rchaud on Friday January 10, 2020 @01:37PM (#22013676)
SEO is only half the issue. Even a great SEO won't be able to get your site to rank if it only has 10 pages and new content is rarely added (think of a small company's marketing website).

Search engines think volume of content and frequency of updates are positive signals, so in the past decade we have seen a deluge of "content marketing", with corporate blogs and media sites chock-full of absolute garbage clickbait and listicles, written by freelance copywriters for pennies per word. All to obtain a higher domain ranking, or domain authority.

While users are looking for the needle in the haystack, websites are just making the haystack bigger to capture users who were looking in a different haystack.

Re:Work on these things by lkschubert8 on Thursday January 09, 2020 @09:20AM (#22001369)
Unless the algorithm could actually detect "quality" (instead of poor analogs for it). Then if SEO couldn't really exploit it, SEO would just be making "quality" content.
Re:Work on these things by cpach on Thursday January 09, 2020 @09:21AM (#22001374)
Funny that you mentioned the issue of googling for recipes.

We discussed it in another thread just the other week: https://news.ycombinator.com/item?id=21938261

I agree that it’s a tricky and quite big problem.

Re:Work on these things by rchaud on Friday January 10, 2020 @01:40PM (#22013713)
It's a problem that's solved by buying a dedicated recipe cookbook. I don't mean to be blunt but at this point, we all know the game these recipe blogs are playing. There is simply no way to monetize users of such basic text content without plastering those pages with autoplaying videos, email signup modals and a million HTTP requests to ad and analytics servers.
Re:Work on these things by throwno on Thursday January 09, 2020 @09:35AM (#22001483)
Not exactly what you're describing, but try this:

https://millionshort.com/

Re:Work on these things by rchaud on Friday January 10, 2020 @01:31PM (#22013574)
millionshort.com is a search engine that lets you exclude high-ranking, SEO-gamed domains from your search. You can remove the Top 10, 100, or 1000 domains from your results. It might not exclude the recipe blogs you're talking about (I dislike those as well), but is a step towards finding more long-tail search results.