Is cloud computing a trap?

I would have not used the same words, but I share Richard Stallman’s opinion over the ‘cloud computing’ hype.’  Whoever has dealt with Richard Stallman knows that he talks straight and direct, but he can do that because he is the original visionary. Putting your data in somebody else’s hands is not smart. It’s not smart to use software on your computer saving your own data in crippled unknown, broken, patented formats and it’s not smart to hand your data to somebody else that you cannot control in any way. RMS to The Guardian:

“One reason you should not use web applications to do your computing is that you lose control,” he said. “It’s just as bad as using a proprietary program. Do your own computing on your own computer with your copy of a freedom-respecting program. If you use a proprietary program or somebody else’s web server, you’re defenceless. You’re putty in the hands of whoever developed that software.”

Some things are just too convenient though, like Google Docs or Zoho. Being able to share and work on the same document with colleagues in different places is a very important thing. I wonder if there are similar Free Software tools that are easily installable.

The latest hype puts MobileMe (and Funambol, fwiw) into the ‘cloud’ but I don’t believe it’s the same thing as Gmail or Gdocs.’  Funambol for sure and MobileMe to my knowledge don’t store data instead of you: these services sync data that you already have locally and keep a backup in sync for you. It’s like ftp-ing your backup tgz over night to a server. You still own your data.

I think that the importance of the web services will increase in the future while importance of ‘desktop’ apps will decrease. Mainly because the desktop concept is changing, fading with increasing availability of mobile computing devices (phones, netbooks, wearable systems) that rely on the Net for retrieving data. Cloud computing is not inevitable, if you think of a computer as something that sits in a room. When it comes to mobile the ‘cloud’ is part of the system and it can be a good thing too provided that you don’t lose control. Probably using web services distributed under the Affero GPLv3 is a start.

Disclaimer: I have a gmail account that I use it only to store mailing lists that I read rarely and for all the trash subscriptions to forums I need to monitor for work or services I don’t use but I need to test once.




  1. > The latest hype puts MobileMe (and Funambol, fwiw) into the cloud’
    > but I don’t believe it’s the same thing as Gmail or Gdocs.

    GDocs related, but is an extensible example: If you use Gears, you have your data in the cloud bu also in your local storage…
    If i use a email client for storing my email (imap access to the server), i have data in the cloud but also in my local space…

    Local persistence of “in the cloud” data are one of the keys for mobile ecosystem development… Mobile devices are “disconnected by design”, but they know how to retrieve data from the cloud and store this data and present it to me when i need, also if I cannot reach the Net.

    I think that Stallman’s opinion is mainly about “security and privacy” of your data: with free software I can check what apps do with my data, every single bit of my data… with closed webapps, I completely lose this control. More, if I’m unable to replicate data on local space, I lose also my data. And this is the worse scenario.

    My 2 cents…

  2. I didn’t think of Gears, you’re right. You can also export all files in ODF format and with IMAP/POP3 access you can keep your data locally. I agree with you that the real issue is privacy and security and there is no doubt that Gmail and similar services cannot guarantee it. Plus using them gives google an incredible power.
    The mobile ecosystem is quite complex. I think it’s a place where open standard and free/libre software will play *the* role. I talked also about these issues in Portland, the slides are here (ODP).

  3. Hi Stefano –

    From my recollection, an AGPLv3 project, Clipperz, proposed a framework for an open and fully encrypted user/password manager. Adoption of the recommended the concept of “zero-knowledge”, where the application knows nothing more about a user than the transactional load that the user imposes and the general selections made. Clippers eliminates being able to easily associate any individual with any transaction, creating architectural anonymity. (

    While Google is not sending us the keys to the kingdom, or the related source, we should be pushing for AGPLv3 or a suitable web services oriented license and zero-knowledge usage as offered by Clipperz.

    I use a myriad of web services. I voluntarily trade certain freedoms for conveniences, but it would be nice if the FOSS community could come together to solve this.

    References –

  4. Ciao Stefano,

    sto smaltendo un po’ di post arretrati da leggere, e proprio negli ultimi giorni sono finito a parlare di questo argomento sul blog. Sintetizzando, dirò che sono sostanzialmente sensibile alla questione privacy e storage dei dati, ma credo che la mera sincronizzazione rappresenti uno step “precedente” rispetto a quello del “cloud computing” (per quanto fastidiosamente altisonante sia questo termine): poter accedere i dati da ogni dove nel mondo, sfruttando software estremamente efficaci (come indubbiamente, lode a loro, lo sono quelli di Google), è un valore aggiungo rispetto alla semplice sincronizzazione. Filtri antispam unici, classificazione dei contenuti, la sincronizzazione stessa su più device: dobbiamo quindi necessariamente rinunciare a tutto questo in nome della libertà? Abbiamo forse dovuto rinunciare ad usare sistemi operativi, quando non ce n’erano di liberi? Possibile che non ci siano altre strade?

  5. Giacomo: per come la vedo si puà accedere agli “stessi dati da ogni dove nel mondo” in due modi: avendo un unico repository centrale accessibile via internet oppure tanti repository sincronizzati tra di loro. Io preferisco avere tanti repository sincronizzati e servizi federati, piuttosto che affidare tutto ad un unica entità.

    La presentazione di Stuart Langridge a OSCON contiene spunti interessanti su come si possono avere servizi online federati e dati replicati su server diversi (non necessariamente dipendenti da UbuntuOne)