Hacker News Reader

Comment by mixedbit | Hacker News Reader

Nice UI and nicely formatted output.

Why does it return .zip that needs to be unpacked? To save bandwidth your could just use gzip 'Content-Encoding' end return the format requested by the user, which would be unpacked by the browser.

Returned file name is [documentcyborg.com].zip, it would be nicer if the domain of the requested document was used instead.

View on HN

For zip file it is not just to save bandwidth, it is also to save disk space as we don't know yet how many user will use it, we preferred to be cautious but in the future you are right we can give straight the file once we know how much disk space will be needed to sustain it. For the returned document name, this is up to discussion if a lot of user ask for it we will change it. Thanks for using the apps and for your comment.

You can store it gziped on your disks, just add 'Content-Encoding' and 'Content-Type' headers while returning it and remove .zip extension.

fixed

As a user, if I ask for a DOC I want a DOC. If your service gives me a zip, it's not what I asked for. As a programmer, this is quite simple to solve as others have said how (store gzipped and serve with correct headers).

Understood, we will change that.

I do conversions like this as part of an app, and I never save to disk - I create the file in memory, and return it directly to the HTTP response as a file download. What stack are you running that makes you save to disk? (Because that also would be a deal-breaker for using this on confidential documents, if you ever allow uploads)

For upload we could do file i/o in memory, our stack allow us. But right now if we did it in memory we would run out of memory.

Reply on news.ycombinator.com