Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
The decentralized WeTransfer (github.com/alexsicart)
115 points by alexsicart on Nov 4, 2017 | hide | past | favorite | 47 comments


This is the problem I see with decentralized projects.

The user experience here, is essentially the same as WeTransfer. The only difference is the underlying technologies.

Now, one would have to either blindly trust that the infrastructure is truly decentralized, or have the time and necessary skills to audit the code.

And then they'd have to trust that web server is actually hooking up to the code that was just audited!

What this means is, "decentralized" projects still require a level of trust, even on the behalf of people who understand what decentralization means.

Next is the problem of the "masses". They already are using WeTransfer, or Dropbox for that matter. Why should they switch to this? Because it's decentralized? Now you have to explain to them what that means and why that's beneficial to them, and if you're successful there, they still have to: trust the code (and they only can do that by understanding it) and trust that you're actually deploying that code on your live interface.

Thus, the decentralization community has got a marketing problem. Fortunately for Bitcoin they've got and continue to have massive amounts of free press, press which does not even actually highlight the decentralized aspect of Bitcoin, just the price action. There is no press movement to help catapult "decentralized infrastructure."

I'm just saying, decentralized projects currently only seem to appeal to people who are interested in decentralization. There needs to be a way to move pass this hurdle if there's going to be the much sought after "mass adoption."

The average person doesn't care how something works, they just care that it does work. And plenty solutions already work for them.

I just hope in the end, we aren't just tinkering for the sake of tinkering.


> I just hope in the end, we aren't just tinkering for the sake of tinkering.

I don't think we are. IPFS is one of the few projects that gives me hope that within 10 years the internet won't be completely carved-up and censored by the big tech companies and governments.


I believe we should build distributed projects, seems equally to products already people love, but using great tech as ipfs, to have permanent and unlimited files.


Large, unlimited, and permanent files sounds great to the average person. One nitpick I have with the IPFS crowd is the use of the word 'permanent'. They know exactly what it means, but the connotation for the average person is incorrect (persistent data for the latter, permanent address for the former). The marketing problems run deep.


But for example, wetransfer only allow to upload 2GB files, my proposal unlimited and permanent. I believe people don't need to know it uses blockchain...


That's a feature, not a product. I'm assuming you can transfer larger files through WeChat if you pay?

I believe decentralised systems serve a purpose (especially in fintech), but I'm not sure this solves an actual a problem (marketing it will be 10,000 harder than building it).

Also, as a business open sourcing stuff hardly ever makes sense. Not sure what your goal was with this (cool tech though!) but if it was to set up a business, I don't think that was a great decision.


censorhip? Here in my country, there are webpages blocked(blocking dns). We see there is a lot of censorhip on the internet. IPFS is permanent, anyone can't eliminate that content. You can get access, with lots of gateways.


Recent reports of google docs arbitrarily removing people's access to their documents seem like pretty good marketing for decentralized solutions.


That was a bug. Are you saying decentralized solutions won't ever have bugs which temporarily prevent people from accessing certain files until the bug gets fixed?


If a bug on your end could prevent other people from accessing data then I wouldn't call that decentralised.


The bug was "we blocked data as we habitually do, but we reasoned incorrectly on what to block," not "we blocked data by accident, blocking itself is a bug". Same outcome, radically different rationale.


Comparative marketing doesn't hold anything back. Remember the 'opensource/Linux is a cancer ' mantra ?


how do Dropbox, Google Drive, iCloud Drive, &c. compete? they have similar features, pricing, &c. dex tech doesn't have to be radically different to be competitive, it just has to be good, fast, and have a competitive angle of some kind: perf, price, privacy, &c.

a more interesting angle, imo, is to create an alternative to the GOOG/APPL/MSFT mobile/desktop platforms using free software and dex tech. a lot of the stuff will be the same, people want music, photos, video, social, but I also hope that the p2p architecture will inspire a lot of novel ideas that will only be available inside of our freedom garden.

#heropunch:matrix.org


Because this tech, need to be visible to HUMANS if we want to create the next internet.


"Thus, the decentralization community has got a marketing problem."

Yes, and it's because security is a feature, not a solution.

99% of people (think your Mother, or your non-techie buddy back home) - do not care that much about these things. 'The feature' they care about is 'sharing'. They can already get that robustly via regular services.

So maybe 'it's nice' to have some degree of extra privacy', but not if they actually have to do something, think about it, install something - yada yada.

For a select few, privacy/security is in and of itself a feature, and might be a primary motivation for use/purchase, but it's a small group.

Example: intellectually, I care about the issue, but really, I could care less that Instagram has 3 of my photos, or that FB has a very small number of posts that I've made.

I believe the 'decentralization' movement will happen when:

A) On the consumer side it's easy, obvious, there is a push - and possibly when adoption isn't necessarily driven by the decentralized nature, rather, something else, a 'benefit' that is derived

B) Enterprise. They will definitely not want their data out in the world willy nilly. Think of how much critical data Salesforce has for example. After a few more leaks, it could be that CIO's everywhere demand more 'on premise' and this could be an opening for more decentralized services.


Huh. Does this tool encrypt your files before sending them? If it is like “WeTransfer”, you would expect your files to be available only to people who know the private sharing link.

However, my understanding is that IPFS will publish the file hash to adjacent notes, so unless they are encrypted people can listen for your provider records and look at your (presumably privately shared) files: https://discuss.ipfs.io/t/is-it-possible-to-store-private-ob...


I'm trying to figure out how this would be decentralised. As far as I understand, merely hosting something on IPFS doesn't make it decentralised if there is only one node.

- Where is my file hosted when I receive it?

- Has your server pinned and seeding that file?

- Is there a distributed set of nodes that also host it?


> As far as I understand, merely hosting something on IPFS doesn't make it decentralised if there is only one node.

Doesn't that apply to torrents, too? If 2 or more people don't own (and share) the exact same content, then of course it's not decentralized.


Sure, I guess the difference is, torrents have an inherent mechanism of becoming distributed. Once someone else has a whole/part of the file, they begin seeding immediately (unless stopped of course). In the case of this app, I just receive a download link which is a proxy for pulling something from the network. It would be something else, if the way to get a hold of this file would be through an IPFS node that the user was running.


>It would be something else, if the way to get a hold of this file would be through an IPFS node that the user was running.

I agree, this seems to sort of bypass IPFS on the download side because my mom doesn’t have to have a node running to download the picture I send her.

However, one potential benefit of this service could be to act as a foot-in-the-door to gain a user base ... I could imagine a .dmg /.exe in the future setting up an IPFS node for your average non-techie user. Still, that doesn’t solve the problem of competing with existing & simple centralized solutions with similar features, as noted in other comments.

Glad to see people building projects with IPFS. Nice work.


I'm not really sure what you mean. IPFS works just like torrents. As soon as you receive the file, you start "seeding" it.


Yes it does. But I don't believe this app is using any actual seeding mechanisms of IPFS from a user perspective. Which is why i'm asking about where the nodes would be to make this distributed and decentralised.

My guess is that the server of this project uploads and "pins" this file in IPFS, in which case it would be no different than a centralised solution. If the server goes down, your file is unaccessible.

Also, when you use this app, you aren't doing this through an IPFS node to then be able to seed it. You are just downloading the already hosted file using a HTTP proxy service that can download from IPFS.


> But I don't believe this app is using any actual seeding mechanisms of IPFS from a user perspective

I'm not sure what you mean here. Any user can "seed" any IPFS file they want, they don't need anything from this service.

> in which case it would be no different than a centralised solution

A distributed solution with N=1 isn't the same as a centralized solution. The difference is that distributed solutions can easily increase N, but centralized solutions can't. In this case, you can distribute the file from as many nodes as you control, which makes this service much better than a centralized solution.

> Also, when you use this app, you aren't doing this through an IPFS node to then be able to seed it

Sure, but you can use something like www.eternum.io (or just run your own IPFS node) to fix that. If you don't want to do either, of course you won't have any of the advantages. It's like saying "cars aren't better than couches, they aren't much more comfortable and they don't run if you don't start the engine".


You might be misunderstanding my position. I think IPFS is great and should be used. But this specific project is not really attending to any problems.

It's in the core functionality of IPFS to be able to host a file and the http proxy is available by default by IPFS. So the real problems are distributing and decentralising the file (such as by running an IPFS node) and this project is not making the slightest attempt at that. Without that there is little point in using this service as it is just like any other centralised hosting service, just so happens that the file is on IPFS.


Oh, I see what you mean now, that it's not much better than raw IPFS. In that case, I might agree, although I think UX matters a lot too.


this project is running an ipfs node in your browser

Also, is not like torrents, if disconnect your node, the file disappears!! Read IPFS docs please


> Also, is not like torrents, if disconnect your node, the file disappears

How is it not like torrents? If you're the only seeder and you turn off seeding, does the file not disappear?

> Read IPFS docs please

You're the one who should read them, if you think IPFS is not like torrents.


the server never goes down, because is not like TORRENTS!


If you only send a download link to one person, how would there be multiple people sharing the same content?


It's not the same torrents than ipfs, completely different!! https://ipfs.io/


fully distributed


So the server which hosts the application stores the file I upload, right? As I recall web implementation of js-ipfs isn't ready yet. This project could potentially be much more useful when js-ipfs is mature enough.


At least, I like the name - IPFS (the InterPlanetary File System)


Note that the submission is not about IPFS itself (https://en.wikipedia.org/wiki/InterPlanetary_File_System), but about Decentralized Cloud, which is a project that is built on top of IPSF.


This HN title needs revision.

Otherwise, great link; nice looking project.


Suggestion: "Decentralized-Cloud IPFS file transfer"


Good one


If you're no longer able to edit the submission title yourself, you can contact the mods via the Contact link in the footer and have them change it for you.


Okay thanks a lot :)


Bittorrent protocol is decentralized wetransfer right?


It would not be inaccurate to consider IPFS as an evolution of the Bittorrent protocol.


I wouldn't disagree with that.


Very different protocols!


From the primary IPFS paper [1], which mentions "torrent" 26 times:

"In IPFS, data distribution happens by exchanging blocks with peers using a BitTorrent inspired protocol"

[1] https://github.com/ipfs/ipfs/tree/master/papers/ipfs-cap2pfs


Not really, with bittorent sender and receiving peers need to be online at the same time.


It is quite obvious that for any file transfer system to work both sender and receiver have to be online at the same time. One sends the data and the other receives it.


has anyone thought that I dont want to store other people's files. that I dont want to store my files on someone else's computer? Decentralized is neat until 1 virus spreads in one second. beef it up then roll it out.




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

Search: