Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Please critique my webapp
27 points by jeroen on June 16, 2008 | hide | past | favorite | 53 comments
Please take a look at http://taggl.net/ and tell me what you think. All comments are welcome, don't hold back.

I would also like to know which applications you think are missing and should be added.

There is an example search query on the frontpage, but this seems more suitable:

http://taggl.net/search/?tag=ycombinator+startup+school+2008

Thanks!




Frankly, I don't have the problem you're trying to solve. And I'm not even sure what problem you're trying to solve is. I do know I never woke up thinking - damn, my tags are acting up again.


Folks, read and re-read this comment when you are firing up a shiny new web app.

What problem are people waking up with who can't wait to give you their credit card numbers for you to fix? If that isn't a problem you are solving, it may be a long slog uphill.


Actually, if I'm not completely mistaken this is a rather cool site function, so correct me if I'm wrong, but does this site scrape the sites listed on the main page for the tags you search for then produce a results page?

I think that's what it's doing, and if so that's pretty neat. It's a great way if you're doing a research project for example to gather social data in one location based on tags instead of going to each individual website finding what you're looking for.

As far as critique goes:

I can't say much by way of development since I'm the design ninja on my team. We'll start there:

On the front page I'd make the search function much more prominent since that is the central component to the reason your site exists. For the most part, the idea is great but the execution lacks in placement of your design elements. For example, again on the front page instead of listing all of the sites being scraped, put them on an about page or somewhere out of the way, but still easily accessible so users will know where data is coming from. Your banner should go up top, in tried and tested results from many other sites similar to this.

Secondly, on the search results, I wouldn't overemphasize the ability to refine the search results by way of a tag cloud. perhaps using a secondary form element so users can define themselves what details they want to disseminate, and making the tag cloud a secondary matter. But most importantly, on the matter of search results: those should be obvious; don't make the user scroll after rendering the data they've searched for. This is both redundant and pointless.

Thirdly I'm noticing you're making extraneous use of line-height properties. Don't overdo it. I know your copytext is a little sparse, and that's fine but don't sacrifice real estate on the page that could be used on it's own page just to make a point.

Overall, I would reduce the front page to maybe just the banner, and the search form (just like Google and even the Twitter search App Summize), with links at the bottom to your blog, contact and about pages, as you have now. Make your results page more readable and more relevant to the information being presented, and utilize screen space more effectively by reconsidering where some of your content elements are.

But, for what it's worth I think this is a great idea for a site, the design isn't bad. I can see you know how to do it, now just work on how to make it work for the user. Just because you know how to navigate doesn't mean the rest of us do.

:) Good luck, and keep reading this thread. There have been some great suggestions so far from the rest of the community.


Your suggestions are very good.

But, as a design ninja, you really think the design isn't bad? I was with you till that line. I thought the four paragraphs above it were on the money and pretty strongly imply the design is bad.

The idea is good, and the design will be good too after two or three more iterations. But right now, that design is bad.


Having a sense of design is one thing, execution when it comes to user impression and interaction is another, which is why you'll (almost, read it again, almost) never see a designer for a magazine or newspaper layout working as a web designer.

I can see he knows how to use the tools he has, and the technologies required, and for that he's done well. But it just needs improvement that only comes from practice, practice and practice. My position isn't changed, it's not BAD by any means, it just needs to be more effective.


I woke up this morning with a very itchy and painful case of tags. I must have been surfing a particularly seedy corner of Web 2.0 last night.


If I'm looking for stuff about a topic in general I could see this as being useful.

I sometimes use del.icio.us as a poor mans search engine to find interesting but somewhat more obscure stuff.


You really need to sanitize that input. It's really simple to inject arbitrary html into your page... best case it messes up the output and just looks bad. Worst case it exposes something etc etc

Not to mention things like this:

http://taggl.net/search/?tag=--%3E%3Cscript%3Ealert(0)%3C%2Fscript%3E%3C!--

http://taggl.net/search/?tag=%2F%22

It doesn't look like you're santizing at all, so someone could have fun with your backend/sql/etc etc


If I had to guess, I'd say he is sanitizing before it hits the database. Not cleaning it up before displaying it to the user, though.


Exactly. My bad, will fix that tomorrow. (it's late over here)


Little Bobby Tables, we call him.

http://xkcd.com/327/


Too slow.

Also, I get the tag cloud thing, but there are just too many to process. There has to be a better UI in there somewhere--something like displaying the now biggest ones first and then asking for more.


The slowness can be well justify here, given the early startup stage the lack of powerful server to drive the search is acceptable. Search requires considerable Disk+RAM power.

Overall the idea is very good!!!


Certainly speed requires fast servers, but if there's not much demand, there's a good chance that it will be faster than a popular service on a single machine.

For example, I bet twitter on a VPS had better uptime than the current version.


I would suggest moving the tag cloud below the results. After searching, the first thing I want to do is see the results. The second thing I _may_ want to do is refine my search. In fact, put it under the pagination links. Also, only showing the top X related/refining tags, and providing a distinct link to 'Show more tags', or something like that, would improve the experience as well, in my opinion.

I would also like to ignore certain services in my search. This functionality may be available to registered users -- I didn't register (sorry!).

Cool idea, though. Would be interesting to see it also search Twitter's unofficial tags (denoted by octothorpes -- '#').


I quite like it. I would restrict the pop-up/preview images to appear only when the user hovers over the link and not when they are in-line with the link - it gets pretty annoying very quickly.


I was trying to create a bigger target, but I can see what you mean. Better now?


I doesn't seem to have changed. I meant to have the pop-ups/previews only when you hover over the actual blue link and not when your mouse goes over the whole grey box background around the link. That way the user hovers over the link and can preview it before clicking without having the pop-ups everywhere.

The pop-ups seem to come up in the same place and if one comes up before another, which is a problem caused by having the pop-ups come up when you go anywhere near the grey bar, it becomes difficult to know which pop-up is for which link. Personally, I would have the user hover over the blue link, then have the grey background box come up as a sort of guide for the eye to follow towards the pop-up, and have the pop-up surrounded with the grey background box as well (at the minute it has no background 'frame'). This way you guide the user from the link to the pop-up and they know exactly which pop-up relates to the link.

Also, I would reduce the size of the tag cloud. I just searched for Adobe AIR and the cloud was huge. Perhaps have the top/most occuring tags and the ability to expand the cloud if you want to.


Looking at the source code:

- minimize CSS & JS (especially the 124kb prototype.js file). You are using 257kb ONLY for javascript.

- gzip

- Why load each company logo on the top separate? put them all in one image file and save HTTP requests

- Unicode error in line 6: "copyright � 2007 J2Solutions" therefore not valid XHTML


You could hide the tags by default and only show them when it's needed. They're quite annoying.


I couldn't agree more, I searched for "dog" and I had to scroll and scroll and scroll to get to my results.

Also, the line "Tags used to be all over the place", trips me up when I scan/read it.


Better, do not show the most irrelevant tags (those in the smallest font size) at all - no one will look through three screens of small text anyway.

P.S. I realized there are actually some search results below only after adding a third tag to the query...


That little "free" image is ostentatious, obnoxious, and silly.


Exactly my first thought. Funny how little things get noticed first.


There's an escaping issue: if I type '\', the search box comes back with '\\'.

edit: similar issue with the single quote (')


Ah, a PHP site.


For me, the ability to quickly search all my tagged stuff from all the services I use is the most compelling part.

Consider just focusing on that. Have a short video that shows exactly how it works on the left. And on the right have a box that allow users to input their credentials, rss feeds, whatever you need, for all their web service and then begin searching all those services with a searchbox that auto-completes as they type. Allow them to do all that WITHOUT having to register. Store a cookie that connects that user to all the info they just inputted which is in your DB. Then tell them to register to save that info. Why require them to register first? If they like your service, and put all that time into adding their info, they will register.


This contains a spectacularly good idea. Registration is a huge barrier. If you can suck people into using your more advanced features without requiring registration, or pushing it as far back as possible, thats a thousand times better. Most people will not register even if they would use it otherwise.


Since searching for tags is what the site is all about make the search box much bigger and put in in the blue section. The size and contrast will attract the eye much better that it does now. Keep the size of the search button consistent with the textbox.

Also get rid of the buoy. I get...help, but for the amount of help it gives me it is annoying. Just type the help text in a smaller size above or below the search box.

Other than that (and the free icon, but that was already brought up) it looks pretty good on the first page at least. The search page really needs some better filtering of data.

Best of luck.


Say what taggl is on the first page rather than "click here to find out what taggle does". That's my two cents.


I would also add a tutorial vid or a video description. I watch those before i make a decision on using a site or not. My two cents as well. That makes 4 cents


I searched for "photography," but I was given a tag cloud of things. Why did I not just get results for "Photography"? Did I just miss them?

EDIT: Yes, I did. They appeared WAYYY down below. Please move the tag cloud under the more important results.... or let the user refine the search with a smaller (maybe it could expand on click) "refine search tag cloud"


Very nice and useful idea. I may actually come back and use it (which happens very rarely with new apps nowadays).

Problems:

Way too slow. Spend more on hosting :(

The "meat" is in the flickr, youtube etc. search results below. The tag cloud should appear blow it IMO. Also, limit its size, and let users enlarge, sort it etc.

How about letting me filter results to just some of the sites you're crawling?


"Way too slow. Spend more on hosting :("

I do hope you're being sarcastic there. Blindly throwing money at things is probably what got twitter where they are now.


Judging by the fact that until they came under open and public criticism they had 3 mysql servers running and a very small development team the only thing twitter was doing wrong was not spending their money.


Not being sarcastic at all!

If a website is too slow, it has a significant adverse effect on my likelihood of using it.

And it's not just me - see Google's assertion that speed is "just about the most important concern of users" - http://battellemedia.com/archives/003076.php


I like the preview feature (when you scroll on a link and it displays the media)

I would work on communicating your idea to the user. What is this website about? What problem are you trying to solve (like others have said)? I wasn't entirely sure until after 4-5 minutes. Many people will probably spend less time trying to figure stuff out.

Finally, from a design perspective, the website looks kind of bare. Splash it with complementary, warm colors (instead of having white space and standard formats for hyperlinks all around). Users will be more inclined to explore the rest of the site that way.


the flickr preview functionality thing is a little wonky. probably not buggy, but the way it works is weird, especially if you fly over all of the links with your mouse.

since the search functionality is the core of your app, you should focus more on the search box. its a little hidden inside things. put the search functionality inside the central blue bar on the main page, google style. make it clear what the users are supposed to do, and then leave the supplemental register/login stuff where it is. imo, that is.

since its tag-centric, perhaps integration with technorati?


Technorati would be nice, but their api does not return tags, which means that I can't create a cloud of related tags.

The public search (the search box on the main page) is only half of the available functionality. For the other half you have to register and login. I appreciate your suggestion to move the search box and I'll seriously consider it, but I fear it will put too much focus on the public search.


could you use the restful interface that they have a la their feeds?

http://technorati.com/tag/ycombinator+2007 http://feeds.technorati.com/tag/ycombinator+2007

etc..

well, if you want to highlight both the search and the register, make them both stand out. i just don't feel like my eyes are attracted to them. i don't immediately know what i'm supposed to be doing on your page when i first load it up. i have to go look and read, and thats a problem, imo. where they are looks more de-emphasized, like supporting text, to me.


My guess is it's so slow because he is pinging the web services realtime for results.

See if you can index them and cache them.

Also:

1. Make the tag cloud have a max of 10-15 tags and have a show more link, make sure the beginning of the results are above the fold.

2. Mix up your results better, and have a sort by feature (sort by flickr, youtube, w/e)

3. Vimeo support would be awesome, digg & mixx support as well, oh and why not add technorati.

4. If you could add Yelp + Maps support I would probably use this to look up cool things to do in any given city.


I like it. Couple of comments:

1) search seems slow to me. What are you using for indexing?

2) Limit the results and display larger tags toward the top - search is about displaying the most relevant


I would put the "separate multiple tags by spaces" right beside or under the search box rather than having to click the little life preserver.

Or why not leave the message out altogether and unclutter the interface. If the user enters commas you can just convert them to spaces on the fly.


Thanks for all the comments. I will respond to more of them tomorrow. For now, I'm calling it a day.


Some comments:

1. I am getting some strpos errors 2. I don't see the benefit of registering. 3. I think that the tag line can be improved....maybe "tag searcher" or "search the web's tags" 4. I don't like the "?" icon next to the Search button. Either replace it, or use it as an Ajax loader :)

-tal


right off the bat, where tags meet did not make sense to me, i get the concept but that sentence needs to be clearer.

then i did a search for "hoodie" and was returned a whole list of tags, I didn't know what to do next. Then I scrolled down and saw the public tags from all the social sites, and clicked around a bit. I think these are the most relevant results. Then the second tag cloud could appear. Also it needs to highlight or emphasize the relevant tags, staring at a huge cloud of world is not intuitive and forces me to figure out what do do next. The concept has legs and could be very useful, U/I needs work to make it easy and useful.


Nice, I dont think it will solve some problem for most of the users. Better to think of it as a new tool for online research. Lacking advanced features, but I'm sure they will come.


Fatal error: Maximum execution time of 30 seconds exceeded in /var/www/taggl.net/www.taggl.net/_data/all.api.inc on line 223


What's the problem that you're solving?


There is a weird font issue with Opera on Linux with your site: all fonts are not anti-aliased...


the tag cloud is ok if its on the right hand side of the screen, and the results displayed in the remaining part of the screen. And like others already said, its pretty slow.


it's slow, Fatal error: Maximum execution time of 30 seconds exceeded in /var/www/taggl.net/www.taggl.net/_data/all.api.inc on line 223


I am not really interested in tagging, so I will limit myself to design.

In addition to the other comments, I think the inconsistency between sometimes using capital letters at the start of sentences and sometimes not is grating. It seems unintentional

The blue bar about halfway down my screen is ugly and not in the right proportion. Go read about golden ratios and ancient temple designs and fix it. Better yet, hire someone who understands that stuff. It's a kinda subtle thing thats the difference between the gateway store and the apple store.

The entire front page should fit in one browser window without scrolling, and the list of supported services should be near the bottom, not top. Your users don't care about other cool start ups. Get rid of that clutter, and put the supported services on the bottom. Then, get rid of tags used to be all over, now they are all here. It's redundant with where tags meet. Are they here to use your app or read your second rate copy?

You should not have Search, register, and login as three things on a list to do. You are search. Collapse the other two into login, register, and get it out of the way. Wufoo does it right, if I remember. You could collapse add apps and comments in the three column part at the bottom, and put something more compelling.

I think it looks like what you are trying to do is a little bit cool and potentially useful to some people but that design and the words on you page bothers me a lot. It's subtle stuff but it adds up to give me a headache. You owe me an aspirin. I don't think the person who made that page is the same one who should fix it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: