Interarchy: A Love Letter

I’ve always thought of Interarchy as the hacker’s FTP client. It has serious street cred, it’s written in Haskell and it started out life as the “Anarchie” FTP client.  However, after becoming a core-contributor to Namecoin, I find myself stopping short of endorsing it as the hacker’s FTP client of choice: Interarchy is not open source.

The frustrating part of being an Interarchy fan is its relative lack of market share. I’ve tried Flow, Fetch, Forklift, Fugu, FileZilla, FireFTP, Transmit, Cyberduck, CrossFTP, and MacFusion; none of them can hold a candle to Interarchy’s flame.

What really sets Interarchy apart is that it doesn’t try to turn FTP into an extension of the desktop.  Interarchy is a finely crafted tool that surfaces the underlying strengths of the protocols (FTP, SSH, WebDav, S3, and even HTTP) using a great user interface. Interarchy just out-innovates all of the other clients combined.

Despite being a superior product, Interarchy’s market share remains stubbornly low.  Even being a favorite of the King Apple Fanboy himself hasn’t done much for Interarchy’s install base.  It’s clear that the great coding going on behind the scenes is being overwhelmed by an inability to market itself.

Sadly, Interarchy’s progress has languished in the past few versions. It’s developed by a single developer, which means there is no need to whisper Shibboleet in support emails but Interarchy can go a long time between updates.  Being a single developer without a large community has meant that the innovative SSH-based plugin system has gone unused.

And then there are minor bugs, like support for minor variations on APIs.  CloudStack, a scrappy competitor to OpenStack, is used by my favorite cloud vendor to power it’s VM’s and S3 clone.  However, for whatever reason, the current version of Interarchy doesn’t support its version of S3 and the developer won’t have time to get around to fixing it until the next major version.

All of this makes me think that Interachy would benefit from a hybrid open-source model.  Currently, the best OSS FTP client is Cyberduck, which isn’t saying much.  The interface sucks, there is no filesystem level integration (in case I want to mount my S3 store as a local disk to help my IDE out), and it is not nearly as powerful as the more vanilla shareware FTP clients.

The situation reminds me of TextMate, the de-facto text editor for OS X.  It started out as a rockstar shareware product that fell behind as the author struggled to keep up.  It’s open-source version was an attempt to help distribute development, but it has sadly remained stagnant.  I feel that the authors of Textmate should have kept a shareware version to support themselves while opening up the source of the core version.

Whereas Textmate ran into a dead-end, IDE’s like IntelliJ, Komodo, and others have all open-sourced  core versions of their previously closed-source product.  The trend for developer tools to provide an open source version is the recognition that these serve as on-ramps for end-users and enables the vendor to capture contributions from outside developers.

The wall separating shareware and donation-ware is already paper thin.  It’s not like Interarchy cannot be had for free; like all shareware it’s serialization process has been thoroughly cracked and can be had without much trouble.  Secondly, as there is no high-quality OSS FTP client for OS X and a free version of Interarchy would undercut the market for Flow, Forklift, Fetch and the like.  OS X’s code signing and app-store process provide a ready-made feature point to differentiate the paid and OSS versions.

Most importantly, however, going open-source means that vendors like GreenQloud and others looking to make it easier to connect to their system would put money into things like compatibility for their APIs.  A core OSS version of Interarchy would also mean Linux and Windows ports, simultaneously expanding the market size for the full-blown version.  The resulting market would be much larger than the OS X only market of the current closed-source product.

Then there is the audibility of the code.  I trust the author not to sneak in nefarious code and I recognize that just publishing your source does magically produce high-quality code audits.  However, now that I am part of what could turn into core internet infrastructure, closed source anything (like the baseband OS in my phone) makes me uncomfortable.  The Snowden revelations prove that nation-states will always be clamoring to crack the security of core infrastructure and political dissidents alike.  The need for high-quality, high-security OSS software with great UX more important than ever.  As a bonus, Interarchy’s Haskell codebase is inherently more amenable to high-reliability software engineering.

Running an OSS project is not easy; community management is hard work!  However, open-sourcing Interarchy would bring in the extra coding talent needed to really take advantage of Interarchy’s abstractions, expand it’s market share, and maybe even get it ported over to Linux and Windows. Finally, they can always retreat to a closed source version in the next release cycle.  But I say plant a seed and see what grows!  Release it under the GPLv3, save the newest features for the paid version, charge for all app-store purchases, and ration 1:1 tech support to paying customers.

Powered by WordPress. Designed by WooThemes