Companies today are facing a perfect storm in information
As the world becomes more and more technological, more and more
of a company’s success is driven – or limited – by its ability
to innovate in IT. And IT innovation, in practice, means
making software do new things.
But companies still developing with 1990s and even 1970s era
software technology are bogged down in software
change management, bug repair, and painful limitations to what
their systems can do.
There is a global shortage of skilled software developers, so
most teams are spread thinly. Even the best developers spend
an average of 50% of their time just dealing with bugs and
defects, and much of the remainder on repetitive tasks, not
just building new capabilities. As a result, crucial software
cannot be deployed – it’s not done when needed, or not
reliable when updated.
Some companies have tried just giving up and outsourcing their
IT, but found that buying commodity software means having only
the same capabilities as one’s competitors – a recipe for
business failure, keeping the CIO from ever adding real value
to the company. Many have decided that’s a dead end. Companies
are demanding more, not less, power from their IT departments
– and just when IT leaders are facing both a capacity shortage
and a quality shortage.
And it’s no longer possible to solve problems by just buying a
faster computer: processor hardware has “hit the wall” with
respect to raw speed. It simply takes too much electricity
(think battery life and heat) to speed up silicon past about 4 GHz.
So hardware makers several years ago switched their focus to
multicore and cluster/ cloud architectures instead. Today
racks of 16-core servers and even 4-core mobile phones are
normal. And chip makers are excited to tell us about their 80-core
and even 512-core machines. “Thanks for nothing,” say many app
developers, because unfortunately traditional programming
languages assume that computers do just one thing at a time.
This means the way people are used to writing software
just doesn’t work on modern hardware, and a powerhouse machine
is often reduced to using just one core – or crashing when the
developer tries to use the other cores. (HTTP Error messages,
anyone?) This limitation couldn’t have shown up at a
worse time for IT departments. Perfect storm, indeed.
A solution 21 years in the making
Fortunately there is a new way to make software: Functional
Programming with Haskell. Haskell, as you may know, was
created over 21 years by a consortium of top universities plus
a broad community of other bright innovators. In
extensive interviews we’ve conducted at FP Complete, companies
say Haskell brings them:
- Productive innovation, needing less time and
fewer people per project
- Quality software that reliably works and can
be reliably maintained
- Scale to handle any required size and speed of
The computer science work that’s gone into Haskell is nothing short
of amazing. That said, I’d give the community as much credit
again for its openness, supportiveness, and willingness to
fold in great ideas received from every direction. The result
has been a language that combines benefits people thought
almost impossible to bring together:
- High-level maintainable code
- High performance comparable to the fastest languages
- Extraordinary bug-proofing with powerful analysis built in to
- Portable cross-platform code
- Ability to take advantage of the most powerful modern hardware
including multicores, GPUs, clusters, and clouds
Beyond just fixing the crisis
And just in time to jump on a number of new opportunities. As
computing systems become more powerful, Big Data has provided
more and more opportunities to innovate. Imagine tracking
every condition that changes demand for your products, and
adjusting pricing, production, and inputs continuously. Imagine
being able to identify complex financial patterns, and to turn
the dial consciously between maximum profit and minimum risk.
Imagine understanding online user behavior and continuously
tuning what offers and articles users see – and being able
to change these formulas at any time without breaking
anything. Imagine analyzing everything a mobile device senses
and proactively giving relevant information to the person
carrying it. And imagine doing all this with no ceiling on the
power of software or hardware that can be applied.
These are the sorts of opportunities commercial Haskell users
are tackling. Some are turning to Haskell to build new
software; more are using Haskell to add to existing systems
that need to do more. Companies that believe their competitive
advantage lies in IT – being able to do things their
competitors cannot match – are choosing Haskell for
high-productivity, reliable innovation.
Our next steps
Over the next three months we will be publishing several detailed
case studies to share what we’ve learned from specific
companies using Haskell. We hope you’ll find these useful in
your own companies. (Feel free to send me your own
success stories and suggestions by emailing aaron /at/
fpcomplete /dot/ com
.) We will also be releasing a major
new online facility for teaching and learning Haskell, as a
free service to the community.
We are hiring salespeople to spread the word about Haskell;
gathering software engineers to deliver commercial tools and
components that open-source projects aren’t quite covering;
underwriting some focused contributions to open-source Haskell
projects; and working on new ways to release and deliver reliable
Haskell distributions. Through sales of tools, training, and
focused consulting, our work will be financially
self-sustaining and we’ll keep contributing to the Haskell
ecosystem for many years to come. You’ll be seeing work from
us on every one of these areas in the coming months.
Companies, researchers, and open-source contributors: we all are
committed to building on the remarkable success Haskell is
already achieving – and to using this great FP system to help
solve the global crisis in software productivity, scale,
Subscribe to our blog via email
Email subscriptions come from our Atom feed and are handled by Blogtrottr. You will only receive notifications of blog posts, and can unsubscribe any time.
Do you like this blog post and need help with DevOps, Rust or functional programming? Contact us.