Minutes of the Baltimore meeting on the Future of VLBI

Participants
--------------------
David Boboltz (USNO)
Ruth Daly (Penn State)
Alan Fey (USNO)
Jack Gallimore (Bucknell/STScI)
Ed Himwich (GSFC)
Ken Johnston (USNO)
Kerry Kingham (USNO)
Julian Krolik (JHU)
Joseph Lazio (NRL)
Chopo Ma (GSFC)
Julie McEnery (GSFC)
Neal Miller (GSFC)
Warren Moos (JHU)
Bob Woodruff (Lockheed) 
Chris O'Dea  (STScI)

Preamble

Chris gave an introduction to the meeting (using Greg Taylor's
viewgraphs), followed by a presentation by Kerry Kingham on the 
Mark5 recording system. Notes were taken by Chris and Jack. 

Summary of the meeting:

Some future projects for an upgraded VLBA
1. Imaging the event horizon of SagA* at sub mm wavelengths.
2. Proper motion of nearby galaxies, e.g., M31
3. Dynamics of the Galactic Center using masers on stars. 

There is strong support for the science that can be achieved via
astrometry and geodesy. 
Astrometry & reference frame studies are important. Stellar VLBI
is severely limited by sensitivity.
Phase referencing has pushed VLBA to fainter sources; but there's still a
lot of science to be done there.

There is still a lot of work to be done on scattering. There is
possible high latitude ionized hydrogen? High latitude quasars are 
showing more scattering than expected. Need better low freq sensitivity 
and new frequencies (74, 150 MHz) for scattering / scintillation experiments.

There's a lot of serendipity in VLBI. Fomalont's "speed of gravity"
experiment, for example. Using water vapor masers to probe black holes ---
no one would have predicted that 10/20 years ago.

When we can detect stars, it will open a new set of astrophysical
questions, strengthen the optical-radio reference frame, and
provide complementarity with SIM. 
The NMA plus wider bandwidth is needed to give the sensitivity to 
do this. 

Too much structure is missing from existing VLBA images. We need
to add the NMA and VLA spacings to make images which sample structure
from 1 arcsec to 1 milliarcsec. This will allow studies of complex
regions (e.g., star formation regions). It will also produce beautiful 
images (e.g., as HST does) which will be useful for public outreach. 

A more southerly dish (Mexico?) is needed to do improved observations
of southerly source, e.g., the Galactic Center. 

It would also help to have broader spectral range (better frequency
flexibility) to get more radio spectral lines.

AIPS has been neglected because of expectations that AIPS++ would
supersede it. But now AIPS needs to receive additional support
since many people depend on it for their research. 

Some complaints expressed how VLBI software in AIPS was really added
as an afterthought, and there's been no real development and progress
researching and implementing new techniques (contrast: recent efforts to
improve deconvolution algorithms vs. no developments of improved, friendly
fringe fitting software.)

RFI will be an increasing problem in the future. More should be done
to deal with it.  

The Health of the Community:
--------------------------
NRAO provides everything an astronomer would want to perform radio 
astronomy, except salary or salary support for students. It's not clear 
that the message is getting out to the broader community.

Outreach is critical and shouldn't be done as an astronomer's hobby. A real
professional effort (like Chandra and HST) should move forward. It is
important to target high school and college students to inspire future radio
astronomers.

NRAO is making progress at outreach, but needs to do more. There should
be a significant fraction of the NRAO budget (e.g., ~3-5%) allocated 
for public outreach. The following are needed: An improved web presence, 
better and more frequent press releases, and more education of 
astronomers on the assistance NRAO provides to novice VLBA users. 
There is an impression that JIVE is doing a good job reaching out to
non-traditional VLBI users in Europe. 

The imbalance between funding by NASA and NSF is a problem. People go
where the funding is. 

Increased funding for university-based VLBI research is very important 
to the health of the field. We need to get more graduate students 
involved.

There are a lot of great VLBI projects which could be done now, but which
are not being done due to a lack of people. 

Regarding surveys of VLBI users:
  a.. It's important to include stellar astronomers in the US VLBI survey.
  b.. It's important to include astronomers who are not already black-belt
VLBI users.
  c.. It's important to get the opinions of young astronomers / VLBI'ers.

The Mark5 Recording system
---------------------------
Advantages of Mark V disks system:
(1)   Disks are cheaper.
a.       Intrinsically cheaper hardware
b.      Cheaper maintenance
c.       Not necessary to have someone on-hand to monitor the recording
(2)   Mark V disks are faster than tapes: ~ 1 Gbit / s recording system.
(3)   Mark V uses commercial technology.
a.       Reduces costs
b.      Naturally improves with quality of hardware market (disk size; i/o
hardware capabilities, etc.)
(4)   Mark V produces better quality output (only one A/D conversion
necessary, fewer errors).

Questions: (all questions and answers here and below are paraphrased)
Q: What is the cost to outfit the VLBA with Mark V recorders?
A: (after some consideration) of order $750,000, but one must also take into
account Mark Vs require less maintenance, materials, operator time, etc.

C(omment): Mark V bandwidth would improve sensitivity by a factor of 2.
A: Mark V is currently the only way to get 1 Gbit/sec recording speeds.

Q: What other systems are being developed, besides Mark V?
A: (1) Thin film heads for tape drives, which will improve recording rate
but not the other issues with tapes (tape breaks, machine maintenance,
etc.). (2) Digital Linear Tape (DLT), but there is currently poor support
for high data rates.

C: (McEnery) We used disk arrays but ran into latency problems.
A: Mark V uses a special (but commercially available) I/O card to spray data
onto disks with no latency issues.

In summary, going to a disk-based recording system seems likely to
be cheaper and more reliable in the long run and will allow enhanced
capabilities.

- end of minutes