Hacker News new | past | comments | ask | show | jobs | submit login
Photo uploading is broken (filepicker.io)
61 points by ananddass on Sept 10, 2012 | hide | past | favorite | 36 comments



I just want to add in my experience in this field and using Filepicker.io as a service:

We spent the better part of two weeks introducing inline images into our social network's stream which worked by the user copy-pasting a URL to a youtube video/vimeo video/imgur gallery page/imgur photo page/direct link to any image mimetype into their status update box and posting it. If they did this, some javascript would (almost) magically show a preview of the attached media and show them that it would be displayed inline. We thought it was awesome and would be the greatest thing ever.

No one used it.

Not only that, our biggest assumption was wrong: Our users didn't know how to upload photos from their computers onto imgur or another photo upload service, if they were told they could copy/paste the URL to an image and have it show up.

Luckily, the weekend after we released this feature Filepicker was announced and we jumped on it right away, I set up a new version of inline images on our site using Filepicker and our previous method combined in less than 2 days, and it was live that Monday, and guess what? While most people still don't add images on our site, the usage rate has almost gone up by 10x.


BidKat actually used a very similar model in our MVP - users would upload images to imgur and then copy and paste the link to add the image to their auction. We had the same results, nobody had any idea how to use it or even what we were talking about.

We looked at Filepicker.io but just couldn't justify the cost. I also had a hard time putting a core part of our business in the hands of somebody else. We see it too often where you take a hard dependency like this and then the service either goes down or worse, gets acqui-hired.

We now just use a simple link the user clicks on to add an image from their computer which then gets processed and uploaded to S3. Took me about 1/2 a day to code up and my monthly cost is about $1.50. I admit it isn't fancy. We don't have drag and drop, you can't add images from Dropbox or Facebook directly, but it works for us and our customers.


We plan to be around. There are investors and customers who would hate to see us end up with the outcomes you listed. So we plan on pushing forward hard on product and market. As regards your comment on pricing...would love to learn about the elements of the pricing structure that was unfavorable to your context. This is valuable feedback. I am adass[at]filepicker.io


Everyone plans to be around :) The fact is that if some company decided to hand you $500 mil to acquire you and shut down Filepicker.io you would (and you should!). In terms of the price, it really came down to the integration. I want everything on my site to look fully integrated. I couldn't justify spending $500 a month to achieve this with Filepicker.io when I could do it myself for $1.50 a month for a good enough solution.


For us the cost is definitely worth it. We're a Fitness social network and there are a lot of groups on the site dedicated to talking about meals, you have no idea how many people use the Filepicker functionality to cross-post instagram pictures of their food.


I have similar experience but from the other end. I have a service similar to imgur. Quite different but for the sake of conversations lets pretend they are the same. You can create content, then save it and get a link to copy-paste into another site.

I have tons of traffic from both the landing page and directly to the creation- and save-page. Most users do save their content for embedding but there is hardly any traffic from the embed-links given to the users. It's not that other people don't click the links, even the thumbnails have comparatively low traffic. This means that most people save their work but they don't know what to do with it later, they simply don't understand why they are given a unique url and that it can be pasted into another site.

Googling for my links also shows that the ratio of published links out of all saved links is very low.


Yeah, imgur makes money because it is basically 99% Reddit traffic. I'm surprised that they are still two different companies...


Oh, how timely! My startup has been pondering whether to incorporate filepicker.io after someone recommended it to us. Photos are extremely important for our site, but for now we've just been using a basic multipart form upload which has been a pain.

I've looked through the docs briefly, but couldn't find exactly this question answered: Can we force our users to crop their photos to a certain dimension with filepicker.io? I played with the Aviary demo and was able to crop, but I couldn't tell if there was a way for the implementor to specify and lock the dimensions.

Currently we have users upload their raw photo, then send them to a cropping page, and then save that, but if we could replace it all in one go with this tool that would be amazing!

Then, if I'm reading it correctly, we're given a URL and have four hours to grab the image from you guys?

Lastly, and I feel a little dirty for asking, but I've been working on this since "what's the worst that can happen?": are any sort of discounts you can offer for a fellow HN user? We're in TechStars Boston and there's a lot of cross-talk among the startups about nifty tools and services to use, so I can assure you we'd promote filepicker.io if it's as amazing and convenient as it seems!


I've been through this precise problem both from a developer's as well as a consumer's standpoint.

As a developer, when we built Unifyphotos(http://export.goyaka.com/) to transfer photos from flickr to facebook, and got some traffic peaks, it was lot of work to scale up the download_from_flickr -> scale_photos -> upload_to_Facebook process, and we had 6 machines just doing this. Wish Filepicker.io was there back then. We could just have worried about the features and not the infrastructure. Even better, if Facebook used Filepicker, there would have been no need for a service to migrate those photos.

From a consumer side, I've often had to download my Facebook profile pic and upload as profile pic on sites like trello, basecamp, etc. If only they used Filepicker.io...


Thanks for the Picplum (https://www.picplum.com) shout-out guys. We had specific requirements around custom UI for using Filepicker. We worked with their team to get it built and integrated in less than a week. Love the support!


To be clear: Filepicker.io isn't broken.


When I read the headline the first thing I thought was "uh oh" ... We're building something with filepicker.io at the moment and I thought the service was down.


Yeh. Totally my bad. should have been more careful with the phrasing. We are not down. All is well :-)


I love that we have to clarify this on days like today


All I want for Photos-Are-Broken Christmas is a plugin to take a Dropbox folder and display a nicely formatted gallery on a Wordpress blog.

It also wouldn't hurt if Dropbox would stop trying to wrap image links in a gallery page. Anyway, here's a picture of a robot's shadow on Mars.

https://photos-3.dropbox.com/thumb/AADVCsDpEacRaKHkyoPhQUU86...

Uploading was trivial. Sharing was a bit harder because they made it harder.

Anyone want to make and sell that Wordpress plugin? It would make sharing trivial for a broad audience with a habit of paying for useful plugins.


You might be interested in OpenPhoto which adds a handful of layers on top of services like Dropbox and S3.

Here's an example of displaying photos from Dropbox (in this example it's S3).

http://current.openphoto.me/photos/list

Disclaimer: I'm a lead dev on the project

https://github.com/photo/frontend


Imgur is beyond simple. http://imgur.com/

Drag and drop or control + V is all you need


Check out my other comment on this thread, but even though you and I think it's incredibly simple almost no one out of my company's (relatively nerdy) target audience could figure it out.


Whenever people on HN and elsewhere say "X is really simple" and someone rebuts them, I like to link to this article: http://www.theatlantic.com/technology/archive/2011/08/crazy-... about how 90% of people apparently don't know how to use ctrl / command-f to find words.

There is no such thing as too simple.


Totally agree with this.

At letitcast, we're dealing with large video uploads (up to 500Mo), image uploads and resumes (pdf). Giving to our users a nice experience was very time consuming.

We had the following requirement for the client side uploader: iE7 support (we have worldwide inexperienced users), progressbar, multiple file chooser, client side file type detection (server side too). We ended up using swfupload + a lot of js. It mostly work but we're not completely satisfied. Last time we checked, there was at least two popular uploaders on Github but both of them are HTML5-only (vs flash based) and didn't offered the same experience in every browser.

We were first uploading directly to the Rails backend but it eventually killed our frontends. We experimented with sinatra async, eventmachine and node.js and endend up using node to handle file uploads, validations, image processing, S3 storage and video encoding (via zencoder). This was our first app with node and in less than a week we had our app in production, handling hundreds of upload/day.

We also had a few suprises with imagemagick causing 100% CPU load (https://www.google.com/webhp?rlz=1C1CHFX_frFR491FR491&so...).

I'm pretty sure I forgot a lot of issues we ran into. We learned a lot along the way but it's nice to see some people are trying to tackle this issue.

FYI the guys at transloadit also work on this.


"You're doing thing wrong!" says person selling new method for doing things.

I think it's slightly perverse that no browser has attempted to offer a method for uploading reduced-size photos. I know it's a new set of patents to trample over but this is a problem that has existed for years that has only been able to use messy plugins to workaround.

Edit: For clarity I'm talking about extending the <file type="input" /> tag to include extra attributes for specifying the maximum dimensions of an image. To me this isn't any weirder than limiting mime-type (which you can through the accept attribute).


You know what is really broken? The "X is broken" record.


"You know what's broken? Things being broken."


After some more digging, I suspect we'll be giving filepicker a try at $work. The value add looks very compelling, especially being able to do things like crop & resize. Our users have been asking for image resize functionality for years, but we've never had the time or priority to do it ourselves.

$600/yr for the pro plan is about one day's developer salary. Rather a no-brainer in my opinion.


I expect that one of the disruptions here will be dropbox. Photo sharing is pretty seamless and even those 'on my machine' photos are pretty easy to export with existing local UX.


Sure. Dropbox is doing an awesome job there. But from the developer perspective, he will still have to maintain multiple integrations into FB, Dropbox etc. If the developer treats dropbox as a local drive folder then he will still have to support things like progress bars, drag and drops etc for local uploads. So the pain will still remain.


Great comment. I can say, anecdotally, that I send photos to friends / family using Dropbox all the time. Most of them are slightly confused by the idea at first, but when the invite e-mail shows up most of them click it and sign up (Dropbox also cleverly incents me to do this by offering more space for referrals).

Dropbox, however, started wanting to copy ALL photos from my cameras and/or Pictures folder to it, which made me bristle because I don't want all my photos, and especially the NSFW ones, on Dropbox.

>But from the developer perspective, he will still have to maintain multiple integrations into FB, Dropbox etc.

Yeah. Especially because I'm averse to FB, but Joe down the street loves it, and Jane mostly uses Picasa, and so on.


I am wondering if you did any testing to determine your pricing? IIRC, patio11 recommends not putting a price for the Enterprise option.


"Photos are a b*tch: Photo uploading is broken"

If you're going to use profanities, then please, at least use profanities.

"In a free society the biggest danger is that you're afraid to the point where you censor yourself." - Tim Robbins


excuse me but these kids are clearly rockstar ninjas and they have earned the right to speak as they choose!


"Brogrammers" more likely.


These kind of ad-hominem comments add nothing to the conversation. Please don't make them on HN.



Why are you posting a link to my comments? Don't be a dick.


These kind of ad-hominem comments add nothing to the conversation. Please don't make them on HN.


>Imagine all the awesome applications possible if you had access to these photos easily.

I cannot. Please, do tell us some of those "awesome applications" if we had access to several billion photos easily:

>One of my favourite services that does more with all these photos is Picplum. Picplum allows you to print and send your photo prints to loved ones.

Awesome... is probably not the right word.




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

Search: