March 2024 ESRFnews

10

INSIGHT

works. Five years later, just half a

dozen ESRF beamlines are still on

SPEC, and even these will all be

running BLISS by 2025.

What is the future for BLISS?

It could be exported to other

synchrotrons. Indeed, it is an open

source software freely available from

the ESRF software repository.

The National Synchrotron Light

Source II at Brookhaven National

Laboratory in the USA has developed

another Python-based system

employing some of the same concepts

as BLISS, and is also on the market. But

BLISS presents several unique features,

such as “Flint”, which allows users to

visualise online data; the BLISS Data

interface, from which any analysis

tool can easily extract online data; and

“Daiquiri”, a framework to build web-

based experimental user interfaces.

Already the BCU is in discussion with

one potential synchrotron client and

there is the possibility that just parts of

BLISS could be exported rather than

the entire system



For more information about BLISS

see A marriage made in BLISS an

integrated beamline control system

M Guijarro et al Synchrotron

Radiation News 2024 doi

1010800894088620232277141

Jon Cartwright

which is vital for making the most of

the EBS’s brilliance.

There are several other benefits.

The data are saved in the common

HDF5 format using the NeXus

standard for metadata, and are

displayed during acquisition, so

that users can make on-the-fly

adjustments. They can also be

accessed online, allowing users

to process them via the ESRF’s

efficient new workflow system,

EWOKS (https://ewoks.esrf.fr),

and close the loop with the running

experiment. This opens up the

possibility of integrating machine

learning algorithms to help optimise

the experimental process (see “AI

revolution”, p12).

When is it available?

On 35 beamlines – remarkably – it

already is. Transitioning to a new

beamline control system is a major

undertaking for any synchrotron

beamlines that have been running

for decades build up an enormous

catalogue of idiosyncratic running

protocols which all have to be

identified rebuilt from scratch

and debugged For this reason

other large research infrastructures

have allocated 10 years for system

upgrades The ESRFs BCU however

took advantage of the EBS shutdown

in 2019 to get a big head start and

demonstrate that their new system

ESRF scientist

Hiram Castillo-

Michel interfaces

with BLISS at the

ESRF’s ID21

X-ray microscopy

beamline. The

new experiments

control system

provides a more

streamlined

experience, and

allows maximum

exploitation of

the EBS.

Why is the experiments control

system so important?

The experiments control system is

the software beamline scientists and

users spend all their time with while

at a beamline. It sets up a beamline

instrument’s hardware, and collects

and displays data. Being universal,

it has to integrate reliably with

everything that is used on a beamline,

provide real-time feedback, and

be relatively easy to understand by

scientists of all backgrounds. It

also has to keep up with the demands

of cutting-edge synchrotron

research – hence the need for an

entirely new system.

What was wrong with the old system?

The old system, SPEC, was developed

by a small US company in the late

1980s, and has been used at the ESRF

from the beginning. For decades

it has been the standard choice for

synchrotrons all over the world, to the

extent that many users and beamline

scientists find typing its commands

second nature. Increasingly, however,

it has been seen as out of date.

Scientists and users have to write

their experimental scripts in a closed,

homemade language – one that does

not readily integrate with others,

notably Python, the lingua franca of

scientific programming. As a result,

it cannot be sufficiently extended

to exploit the full potential of the

ESRF’s new X-ray source.

What is the new system?

The new system is known as BLISS

an homage to the old name

BeamLine Instrument Support

of what is now the ESRFs Beamline

Control Unit BCU In development

at the BCU since 2015 BLISS is

written in Python allowing easy

integration with users own tools and

software libraries Another difference

is that it is not like SPEC limited to

data collection in stepbystep

fashion between experimental

acquisitions but can instead scan

continuously as a base feature This

makes data collection much faster

A new experiments control system has been deployed by the vast majority of ESRF

beamlines, and has major advantages.

Experimental BLISS

E S R F / J E N S M E Y E R