tecosystems

Del.irio.us, Or How Not to Clone a Service

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

Anybody with an interest in how communities react to forking and clones of a project should watch this thread on the del.icio.us discuss mailing list. The background is this: a programmer named Steve Mallet essentially cloned Joshua Schachter’s del.icio.us service, down even to the URI level (his service is called del.irio.us). The passion on both sides of the argument has been illustrative of some of the emotions invested in both the del.icio.us application itself as well open source. The main issues are:

1. Open Source: Part of the reasoning for creating del.irio.us, according to Mallet, was the need for an open source version of the code that runs del.icio.us. To this issue, I’d ask only this: what is the problem that’s being solved by open source here? Development speed and contributions is the most likely answer, and it’s certainly possible that del.irio.us will evolve more quickly than del.icio.us. That said, I’m not sure that’s a given, because of how centrally delivered services operate (yes, I’m aware that del.irio.us can be run locally). As I’ve written many times, open source is a tremendously powerful way to develop software and many if not most projects can benefit from it, but it is most definitely not a development panacea. Also, how long will this remain a competitive differentiator?

2. Social Software: It’s interesting to me that the code behind social-software can come to be as beloved as the community that it supports; as David Smith notes, Shirky has written of this before, but Smith says it best I think:

It’s hardly surprising that some of us who have come to use del.icio.us a great deal might feel just a little affronted, on Joshua’s behalf more than anything else, at such a naked act of appropriation.

Net net, if you were think that community members don’t care too much about the bits, think again. What’s more, potential users or partners of del.irio.us may have to contend with an irritated (and numerically larger) population of del.icio.us users in the future.

3. Degrees of Cloning: A point raised by many in this debate has been the down to the minutiae level of replication in Mallet’s application. This implies an awareness – a healthy awareness, IMO – that some replication can be a very welcome thing, and that cloning is not a binary, black/white issue. Instead it’s the degree to which Mallet cloned del.icio.us that has irked many.

4. Community Division: The most worrisome concern for most posters however is not the technical replication, but rather the potential fracturing of the current user base that could occur if del.irio.us poaches some current del.icio.us users. While this is a legitimate concern, I still believe that playing catchup from a user contributed data perspective is far more challenging than any technical arms race. As a result, it’s my belief that del.icio.us has a very healthy lead on services that clone its functionality; at least until one of them produces a solution that’s head and shoulders better than del.icio.us, which clearly isn’t the case at the moment.

5. Cloning as a Necessity?: One of the interesting defenses Mallet presented for his actions was the need to preserve some semblance of a user experience from del.icio.us to del.irio.us. Here’s what he says:

the interface is a rip off: I suck at CSS & this is the basic layout
that came with the software that runs de.lirio.us. I’m totally open to
changing it up. I do want to keep the same UI though. No sense
screwing up what people know how to do now.

Personally, it’s my opinion that he’s overemphasizing the UI issues (it’s just not that complicated) for the purposes of defense, but I think the point in general is a legitimate one. One would have a difficult time, for example, delivering a completely unique word processor because Microsoft Word has become the standard set of expecations. But what constitues necessary-for-users feature replication, and what is simply gratuitous?

All in all, I think this example highlights the fact that there are right and wrong ways to clone applications or services; or maybe that’s wrong – there are righter and wronger ways if not an absolute black/white determination. In this case, I think the degree to which del.irio.us flatly copied much of the del.icio.us design is not entirely appropriate, but at least in the short term I don’t expect it to make much difference. Nonetheless, for any open source or commercial vendors out there contemplating similar replication I’d urge you to take a look at the concerns mentioned by del.icio.us fans. Some might have been avoided with a bit of community involvement.

9 comments

  1. The fracturing issue is what bothers me. Nice that the dude acknowledges it openly though.

    It'd also be nice if there was a way to post everything you've got at del.icio.us to that so as to give it a nice kick-start.

  2. Thanks for this mention, Stephen. I largely agree with what you're saying. I'll try to say something about it over at my blog later on today.

  3. It doesn't bother me so much. Does there need to be only one of everything? Steve seems to have been very up front about what he was doing. Presumably he wouldn't have gone to the trouble if de.licio.us had been both open source and open content. And presumably no one will switch horses if they don't see a real value in that combination.

  4. David: no problem, will give your blog a look.

    Rob: it's really a matter of degrees to me, Rob. i agree that Mallet's transparency defuse some of the irritation many are likely to feel. nor do i have a problem with people cloning services; competition is a good thing. but what i tried to highlight is the degree to which things are cloned. there's some replication that's necessary from a user perspective, and then – IMO – there's gratuitous copying designed merely to perfectly replicate an environment and make one indistinguishable from another.

    i have no problem with the former, but i believe the degree to which Mallet copied del.icio.us is not appropriate.

    i'll admit, however, that's it an entirely subjective opinion and that there's obvious room for disagreement.

  5. Danno: I could be mistaken, but I thought del.irio.us had an import function.

  6. Stephen,

    Can't get a Trackback through to this post (at least, that's what TypePad's saying at my end of the link), but I have posted here: http://www.preoccupations.org/2005/03/delirious.h….

    Best wishes,

    David

  7. "Part of the reasoning for creating del.irio.us, according to Mallet, was the need for an open source version of the code that runs del.icio.us. To this issue, I'd ask only this: what is the problem that's being solved by open source here? Development speed and contributions is the most likely answer"

    Yes and no. The big differentiator as I see it, is that I don't want absolutely *all* my bookmarks visible to the entire populace. It also puts the control back in your own hands, as you don't have to worry about del.icio.us suddenly being shut down, and whoops, there go all your bookmarks.

  8. The point is, it's a closed service.

    1. While improbable, del.icio.us or Yahoo! can decide anytime (in several years) that it's not worth the hassle anymore and close down the service. Or their server might crash and take everything down, and they decide to stay down. Bummer. Out of luck.

    2. Private bookmarks. I don't want my bookmarks out in the open. Sorry. They are mine.

    3. Extensions. Not very much with del.icio.us. Quite possible with de.lirio.us.

    I agree that it's not very nice to clone an existing project to the core. However, there seemed to be a demand for it, or it wouldn't have happened. Also, the service initially wasn't very much substance in code, or it wouldn't have been possible to clone it that fast.

  9. The del.icio.us thread link is (or has been) taken down. I’d love to read it, if you have an alternative uri maybe? Oh and thanks for the story.

Leave a Reply to Rob Cancel reply

Your email address will not be published. Required fields are marked *