Hacker News Reader

Comment by holler | Hacker News Reader

> Content is either in an endless scroll or it's only accessible through a search box and you have absolutely no idea whether the site doesn't have the content you wanted or if you just failed to use the correct terms.

What alternatives or solutions do you envision for search? How would you build a content heavy site in a way that gives a user confidence when searching for something?

> Now your only option is a web page with embedded pictures which may as well be a paper encyclopedia entry, or a video that may or may not offer you any useful way to navigate it aside from jumping around at random and most certainly doesn't give you rich annotations throughout.

Maybe I'm mentally blocked in thinking of the era of flash intros and marquee tags, but do you have any examples of sites that achieved what you're describing?

Were the annotations created by the site owner or by the community?

View on HN

> What alternatives or solutions do you envision for search? How would you build a content heavy site in a way that gives a user confidence when searching for something?

I think a big difference is the move away from folder-based organisational structures. Once upon a time, web sites were collections of HTML pages in a filesystem, and so a hierarchical organisation was natural; people started moving away from just a filesystem (or FTP server) into a CMS, but many older CMSes had a similar focus on folders as an organisational structure. It worked quite well in the hands of competent users, although there were always those people who didn't quite get it (or maybe just had poor attention to detail) and if left unsupervised they'd turn it into a big mess.

But then along came blogs, and suddenly there was a shift to posting content chronologically, with non-hierarchical tags; and as blog software (such as WordPress) came increasingly to be used as a CMS even for non-blog sites, that approach spread to CMSes as well. And I think it makes it harder to logically organise content, and it results in less focus and time spent on content organisation, and the end result is content becomes much harder to find. Blogging also encouraged an approach where people would create a new page rather than editing an existing one, resulting in an intermittent stream of disorganised piecemeal updates in response to new developments, rather than presenting all the information as a cohesive whole.

I think software (especially documentation) is one of the areas where we have this problem the least, because a lot of software people use tools (Git, Jekyll, etc) which still encourage hierarchical organisation. It is really non-software stuff, for example websites of government agencies, university administrations, corporations, etc, which have gone down-hill the most.

> Maybe I'm mentally blocked in thinking of the era of flash intros and marquee tags, but do you have any examples of sites that achieved what you're describing?

I don't remember many websites like that, but certainly a lot of multimedia CD-ROMs in the 1990s had that kind of design, and while there is no reason from a technological perspective why you couldn't replicate that style on the modern web, it just doesn't seem to be very popular.

The reason multimedia CD-type sites aren't made anymore is because of copyright and licensing costs. You can't store a video on your site that you don't own, but you can embed a Youtube link. Same goes with songs; embed a Spotify link.

It was a much more fluid experience on things like Encarta 97, where MS shelled out the money to license a lot of audio and video. It was completely offline, so there was no possibility of filling the article pages with iframes of content served by third party sites.

> What alternatives or solutions do you envision for search? How would you build a content heavy site in a way that gives a user confidence when searching for something?

At the scale of Netflix? Start with a simple list of all titles in the catalog. One I can search with CTRL+F, or maybe narrow down with an incremental search, if you really have to do some fancy JavaScript in it.

It's not until we get into "Amazon scale" when the simple idea of giving people the whole dataset to look at breaks down.

The kind of operations done against most sites - whether it's a media streaming service or an on-line shop or an e-mail client or a planner - they all fall into a very common set of patterns. Filter by piece of information you know (like title, or genre; both "include matching" and "exclude matching"). Compare a bunch of candidates relative to a set of criteria (e.g. length, price, features X, Y, Z). Compare a candidate against a class in which it's in. Etc. Those are simple and universal patterns for dealing with data. Those are the interactions you want to expose if you actually care about your users achieving their goals.

The way I see it, most content services would've offered a much better UX if they just let you download the catalog as an Excel sheet, so that you can use a tool that actually tries to help you deal with data.

They don't want you to know how few titles they actually have.

There's nothing about old web organization that takes away new web dynamism when added. Just give me a little "collections" link somewhere in a corner or at the bottom. Something I can click on and get everything you wrote on a particular topic. For example, if you're a game review site, let me see a list of every PC game you've reviewed sorted by date. Some sites have that but many don't anymore.

> How would you build a content heavy site in a way that gives a user confidence when searching for something?

Use Good information architecture:

1. Have breadcrumb links at the top of the page so the user can find their bearings quickly.

2. Use categories and tags carefully. Don't add irrelevant or imprecise tags to articles.

3. If your site has over 100 pages of content across multiple categories, consider doing user research to better understand how site users navigate your site and think about categories. This can be done using an exercise called card sorting, where users look at a list of pages (shown as individual index cards), and then organize the cards into categories that make sense to them. This can be done with pens and a handful of index cards, but large sites would probably use an electronic solution like Optimal Workshop to generate tests that can be done online, and crunch the numbers.

Alternatives:

Static, indexable pages. Don't ever, ever build your own search, just use a google query with a site: filter. If that doesn't work, redesign your site.

Examples:

Basically any static page. Yeah, flash and marquee were annoying but you could just turn them off.

Reply on news.ycombinator.com