Well for one thing, I'm human and don't fancy rating my friends &
collegues. At least not so that they know how I rated them.
So, what would happen if the rating system was anonymous..? I would
know who I voted for/rated, but no-one else would know. Of course,
once someone got a bad grade, he/she would remove his/her account from
the system. So after awhile it would be a craftsman-list.
Stripping away text-comments in those ratings would minimize the
posibility of "language recognition". However, that would also bring
up the problem of "what to rate" and "how to interpret the ratings"
instead.
In such an anonymous-vouching-system;
1. What kind of ratings would there be?
2. How would you bring up confidence in the ratings?
3. Would you include language-ratings (Java-skill-level,
Python-skill-level etc.)
4. Would you include technology-ratings (database systems, operating
systems, libraries..)
5. Would you include methodology-ratings (TDD-skill eg.)
6. Would you include social-skills (pairing/communication eg.)
7. Do you think it would work?
--
You received this message because you are subscribed to the Google Groups "software_craftsmanship" group.
To post to this group, send email to software_cr...@googlegroups.com.
To unsubscribe from this group, send email to software_craftsma...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/software_craftsmanship?hl=en.
Is trust transitive? I don't think it is. If A trusts the technical
skills of B, and B trusts the technical skills of C, that does not
mean that A trusts B's evaluation of C's technical skills (unless A
explicitly says so). So in a vouching system system the direct
relations should count the most.
I can trust somebody's technical skills, but that does not mean that I
trust his project management skills. The context of each vouching
should be made explicit.
Some people vouch for every drinking buddy they have, whereas others
vouch for only a few well chosen people (like Ron Jeffries said in
http://www.threeriversinstitute.org/blog/?p=187 "I trust [Kent Beck],
and about three other people, to make good short game decisions"). A
vouching system should try to estimate the importance of each person's
trust.
Given A vouches for B, when B does something stupid, then both B's and
A's reputation should suffer. Or at least the importance of A's trust
should suffer. That way it might be possible to make a difference
between those to vouch for anybody and those who vouch very carefully.
The previous point leads into the need to be able to detect when
somebody does something good or bad. Should and could the vouching
system take into consideration all the projects which the people have
been part of? Is it possible to reliably evaluate whether a project
was a success or a failure, and what was the reason for the failure?
Or is such a measure impossible, like http://martinfowler.com/bliki/CannotMeasureProductivity.html
claims about measuring productivity?
When a client has worked with some developer and has had good results,
then he will trust that developer also in the future. Maybe the client
will also have a measure of trust towards those who the developer
vouches for. But if some of the developer's recommendations is bad,
then the client might not anymore trust the recommendations of that
developer, even though he trusts the developer himself.
So, before discussing the details and logistics of such a system,
could we address the following?
1) What is the goal of a vouching system?
It appears that the 'craftsmanship cooperative' thread switched from a
group of independents who wanted to work with each other to the
assumption that a vouching system is something needed. If you have
people who like to work together and are productive while doing it,
then the need for some sort of 'I vouch for this person' goes away.
2) What is the benefit of a vouching system?
When the idea of vouching or certification or trust comes up, as it
has a few times before, it tends to jump over the initial question of
'why?' So, supposing there was a vouching system, what would you seek
to gain from it? What benefit would it give to you, as a developer
seeking to improve yourself?
3) Why would I vouch for anyone?
Suppose there is a valid answer for the first question, and some sort
of vouching system is worthwhile. Why would I vouch for anyone at all?
As some posters have mentioned, I am putting my own reputation on the
line by vouching for someone publicly. What benefit do I or the
community receive?
Personally, I have vouched for people in private, when recommending
someone for a job, or directing a prospective client to another
person. However, in a public forum, I can't possible think that I
would vouch for a person. That would be completely out of context.
Would a prospective client want to come to a site where someone says
'hey, joe can do exactly what you need,' considering that I have very
little idea, if any, of what they need?
Now, suppose, as the Olof originally asked, we start including
specific things: languages, techniques, processes. What good does this
do for a client? What good does this do for a community member? Would
a client actually know how to look through those categories to find
what they need? As a community member, if I looked through them, what
would I gain?
-Corey
> --
> You received this message because you are subscribed to the Google Groups "software_craftsmanship" group.
> To post to this group, send email to software_cr...@googlegroups.com.
> To unsubscribe from this group, send email to software_craftsma...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/software_craftsmanship?hl=en.
>
>
>
>
--
http://www.coreyhaines.com
The Internet's Premiere source of information about Corey Haines
I guess the goal of such a vouching system would be to be a better
system than certifications has been traditionally. So that
managers/those with the money can have more reliable measure of skill
than before.
Of course with technique-based skill-ratings, the interpreter of the
information would have to be quite skilled him/herself, just to be
able to understand what he/she is reading.
>
> 2) What is the benefit of a vouching system?
> When the idea of vouching or certification or trust comes up, as it
> has a few times before, it tends to jump over the initial question of
> 'why?' So, supposing there was a vouching system, what would you seek
> to gain from it? What benefit would it give to you, as a developer
> seeking to improve yourself?
Reliable certification is an industry problem for sure. So the benefit
would be throughout the industry. Those that want higher quality would
have a place to look.
>
> 3) Why would I vouch for anyone?
> Suppose there is a valid answer for the first question, and some sort
> of vouching system is worthwhile. Why would I vouch for anyone at all?
> As some posters have mentioned, I am putting my own reputation on the
> line by vouching for someone publicly. What benefit do I or the
> community receive?
Well my OP mentioned anonymous vouching. Even so there would be no
direct benefit rating anyone at all.
>
> Personally, I have vouched for people in private, when recommending
> someone for a job, or directing a prospective client to another
> person. However, in a public forum, I can't possible think that I
> would vouch for a person. That would be completely out of context.
> Would a prospective client want to come to a site where someone says
> 'hey, joe can do exactly what you need,' considering that I have very
> little idea, if any, of what they need?
>
> Now, suppose, as the Olof originally asked, we start including
> specific things: languages, techniques, processes. What good does this
> do for a client? What good does this do for a community member? Would
> a client actually know how to look through those categories to find
> what they need? As a community member, if I looked through them, what
> would I gain?
As I mentioned above - a non-technical client would really need
expertise to understand the rating.
On Wed, Jan 20, 2010 at 2:07 PM, Olof Bjarnason
<olof.bj...@gmail.com> wrote:
> 2010/1/20 Corey Haines <corey...@gmail.com>:
>> 1) What is the goal of a vouching system?
>> It appears that the 'craftsmanship cooperative' thread switched from a
>> group of independents who wanted to work with each other to the
>> assumption that a vouching system is something needed. If you have
>> people who like to work together and are productive while doing it,
>> then the need for some sort of 'I vouch for this person' goes away.
>
> I guess the goal of such a vouching system would be to be a better
> system than certifications has been traditionally. So that
> managers/those with the money can have more reliable measure of skill
> than before.
How would a vouching system cause the measure to be more reliable?
Wouldn't the hiring manager have to know the people who were vouching?
>
> Of course with technique-based skill-ratings, the interpreter of the
> information would have to be quite skilled him/herself, just to be
> able to understand what he/she is reading.
That is a point that is often over-looked.
>
>>
>> 2) What is the benefit of a vouching system?
>> When the idea of vouching or certification or trust comes up, as it
>> has a few times before, it tends to jump over the initial question of
>> 'why?' So, supposing there was a vouching system, what would you seek
>> to gain from it? What benefit would it give to you, as a developer
>> seeking to improve yourself?
>
> Reliable certification is an industry problem for sure. So the benefit
> would be throughout the industry. Those that want higher quality would
> have a place to look.
Why is reliable certification an industry problem? How would vouching
ensure 'higher quality?' I wonder how many people out there don't
want 'higher quality'
>> Now, suppose, as the Olof originally asked, we start including
>> specific things: languages, techniques, processes. What good does this
>> do for a client? What good does this do for a community member? Would
>> a client actually know how to look through those categories to find
>> what they need? As a community member, if I looked through them, what
>> would I gain?
>
> As I mentioned above - a non-technical client would really need
> expertise to understand the rating.
So, unless the person was technical, having anything more than just
'I, Corey, vouch for this person' isn't that useful.
Isn't that what Laurent's is?
-Corey
My suggestion was an *anonymous* vouching system, as opposed to a
public one. Be the result of the observation "I don't want to rate
people I know, so that they know how I rated them.".
Microsoft certifications and the like are not worth much at all in my
humble experience.
>
>>
>> Of course with technique-based skill-ratings, the interpreter of the
>> information would have to be quite skilled him/herself, just to be
>> able to understand what he/she is reading.
>
> That is a point that is often over-looked.
>
>>
>>>
>>> 2) What is the benefit of a vouching system?
>>> When the idea of vouching or certification or trust comes up, as it
>>> has a few times before, it tends to jump over the initial question of
>>> 'why?' So, supposing there was a vouching system, what would you seek
>>> to gain from it? What benefit would it give to you, as a developer
>>> seeking to improve yourself?
>>
>> Reliable certification is an industry problem for sure. So the benefit
>> would be throughout the industry. Those that want higher quality would
>> have a place to look.
>
> Why is reliable certification an industry problem? How would vouching
> ensure 'higher quality?' I wonder how many people out there don't
> want 'higher quality'
Probably 90% of the software industry does *not* want higher quality,
but are satisfied with shrink-wrap-crap (in my eyes). My view is also
that we live in an industry thriving even though the general economy
is in a decline, or at least standstill.
But that does not mean there isn't a real need for higher-quality
software, and a market for it. And thus a market for efficient ways to
find software craftsmen, which as I perceive it, aim for
higher-than-average-quality software formulation.
I'm not suggesting an anonymous vouching does ensure higher quality. I
wanted to open up for a discussion if it was an improvement enough to
be of use, compared to public vouching systems.
If the systems mentioned on this list so far (wevouchfor and David
Starrs system, what was it called?) are alread anonymous, I guess this
topic does not bring anything new to the discussion.
>
>>> Now, suppose, as the Olof originally asked, we start including
>>> specific things: languages, techniques, processes. What good does this
>>> do for a client? What good does this do for a community member? Would
>>> a client actually know how to look through those categories to find
>>> what they need? As a community member, if I looked through them, what
>>> would I gain?
>>
>> As I mentioned above - a non-technical client would really need
>> expertise to understand the rating.
>
> So, unless the person was technical, having anything more than just
> 'I, Corey, vouch for this person' isn't that useful.
Nope, unless the system could give some kind of "destilled" or
"reduced" information, readable by non-expertise. I doubt it though.
>
> Isn't that what Laurent's is?
If Laurents' system is an anonymous vouching system, I guess it is?
>
>
> -Corey
>
> --
> http://www.coreyhaines.com
> The Internet's Premiere source of information about Corey Haines
>
--
You received this message because you are subscribed to the Google Groups "software_craftsmanship" group.
To post to this group, send email to software_cr...@googlegroups.com.
To unsubscribe from this group, send email to software_craftsma...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/software_craftsmanship?hl=en.
If that is the goal, then would a solution be a better interview
process?
That goal has the assumption, that currently managers can not reliably
measure the skill of an interview candidate. This leads to the
question - why? What is wrong with the way that technical people are
typically interviewed?
If we would have a list of things that do work and don't work in an
interview, would that help the managers to more reliably measure the
skills? Or is it impossible to measure somebody's skills in an
interview?
If technical skills are required to estimate the technical skills of
someone else, but the company which is hiring people does not have
any, then could the interviewing be done by some consulting company
which has the skills to evaluate the candidates reliably?
Well a better interview process would be for developers to pair with
each interviewee to make a human judgement.
That is something I would do if I were to hire - but it is
time-consuming. And the "promise" of a vouching system would be to
bring down the number of such resource intensive interviews needed to
find what you are looking for.
I'm leaving this topic for the moment - I don't think we have more to
discuss right now.
>
> That goal has the assumption, that currently managers can not reliably
> measure the skill of an interview candidate. This leads to the
> question - why? What is wrong with the way that technical people are
> typically interviewed?
>
> If we would have a list of things that do work and don't work in an
> interview, would that help the managers to more reliably measure the
> skills? Or is it impossible to measure somebody's skills in an
> interview?
>
> If technical skills are required to estimate the technical skills of
> someone else, but the company which is hiring people does not have
> any, then could the interviewing be done by some consulting company
> which has the skills to evaluate the candidates reliably?
>
There's a very important point I see that comes up time and time again
in discussions around vouching, or certification, and that's this:
Corey Haines wrote:
> 1) What is the goal of a vouching system?
To rephrase Corey's question, what are we trying to solve?
As an independent consultant, I always look for new ways to get the word
out there about what I do and who I am.
As a company, like perhaps 8th Light or Obtiva, they may want to be able
to show the caliber of their craftsman.
As a company looking to hire a company, they may want to check to see
the references of the employees working for the company.
As a company wanting to hire developers, they may want to be able to see
whether the claims made by the candidate are backed up with actual
experience.
But it seems like, to me, the focus is always on that last line - the
non-agile organization looking to bring people in. After all, if they
were involved in the community at all, they would already know who many
of us were.
But, none of us here /run/ a company like that. So I don't see how we
can effectively create a system of vouching, recommendations or
certifications that would meet their needs.
--
Cory Foy
http://www.coryfoy.com
http://twitter.com/cory_foy
To rephrase Corey's question, what are we trying to solve?
Corey Haines wrote:
1) What is the goal of a vouching system?
As an independent consultant, I always look for new ways to get the word out there about what I do and who I am.
As a company, like perhaps 8th Light or Obtiva, they may want to be able to show the caliber of their craftsman.
As a company looking to hire a company, they may want to check to see the references of the employees working for the company.
As a company wanting to hire developers, they may want to be able to see whether the claims made by the candidate are backed up with actual experience.
But it seems like, to me, the focus is always on that last line - the non-agile organization looking to bring people in. After all, if they were involved in the community at all, they would already know who many of us were.
But, none of us here /run/ a company like that. So I don't see how we can effectively create a system of vouching, recommendations or certifications that would meet their needs.
... I'm new to this group and very intrigued by the concept of
software craftsmanship. It has the smell of an idea with serious
legs, not to mention sound historical precedence. ("I apprenticed with
Kjetil who builds the best longships in Sognefjord. As a journeyman I
traveled, working with the following masters ...")
James, I love the idea of "ranking" based on how strongly one's
thinking was influenced by whom. I imagine it would give a very useful
indication of what to expect when working with a person. By
comparison, the notion of a ranking system by peers or others seems
oddly terrifying, in a distracting sort of way: What are we trying to
achieve? Work as a team to build better software or try not to be
voted off the island? I think that the notion to be "page ranked"
popularity-contest style would get folks all worked up about the
latter, at the expense of the former. It's understandable because it
might influence peoples' ability to make a living just a little bit
too directly.
Having said that, the "page ranked" popularity-contest style may have
its place: I wonder if one could rank teams instead of individuals? A
common practice in the industry is to throw together people who are
suddenly supposed to work together smoothly to deliver a project.
However, it takes time to build a good team which is truly on top of
its game. It also takes a blend of masters, journeymen and
apprentices: If you have all masters, who is going to challenge them
to explain and question their ideas? A shop full of apprentices with
no one else to ask .... well ....
If there were a ranking (... perhaps with awards, etc....) of entire
teams, it would encourage the practice of giving projects to (...
cross-functional, used to working with each other, etc.) teams instead
of the other way round. You would join a team instead of a project. If
you look at the team members' lines of influence (see above), you
would have a pretty good idea what to expect. The "lines of influence"
would get us into the direction of "lineages" of masters, in the
Eastern sense. Finding a good way to rank entire teams might move us
towards "workshops" along the lines of medieval European craftsman's
guilds, etc.
Interesting times indeed. Love the discussions.
Robert
http://robertblogs.wordpress.com
On Jan 24, 9:58 pm, James Martin <jimmymar...@gmail.com> wrote:
Tried some storyboarding - Whose thinking and experience is influenced
by whom? - http://robertblogs.wordpress.com/2010/01/31/software-craftsmanship-coding-under-the-influence/
.
Robert
On Jan 30, 10:19 am, Fred Ballard <fredb...@gmail.com> wrote:
> I love the idea of ranking based on how strongly one's thinking was
> influenced by whom, too. It's a way to encourage craftsmen to look for
> influences outside themselves and could lead to some interesting
> discussions.
>
> (Mathematicians have their Erdös <goog_1264817918154>
> number<http://www.oakland.edu/enp/readme/>.
> It indicated how close you collaborated with him. Erdös had an Erdös number
> of 0. If you collaborated directly with him you are a 1. If you wrote a
> paper with that person you are a 2, and so on. Erdös was even more
> peripatetic than Corey. Erdös had no home and just contacted a fellow
> mathematician to announce his imminent arrival at the next place in his
> journey, knowing this person would pick him up and take him in. Could we be
> moving towards a Haines number? 8^) )
>
> I think ranking individuals is difficult at best, and could degenerate into
> a popularity contest at worst.
>
> On the other hand, ranking teams does seem easier to do -- their output and
> effect is usually more readily viewed -- but this too could become which
> team is more popular than the other.
>
> Fred
>
> On Sat, Jan 30, 2010 at 8:51 AM, Robert R. <rrep...@gmail.com> wrote:...
Very interesting idea! What would be a measure of "collaborative
distance" among developers .... ?
Tried some storyboarding - Whose thinking and experience is influenced
by whom? - http://robertblogs.wordpress.com/2010/01/31/software-craftsmanship-coding-under-the-influence/
.
...
On Jan 31, 8:43 pm, Robert Reppel <rrep...@gmail.com> wrote:
> Very interesting idea! What would be a measure of "collaborative
> distance" among developers .... ?
>
> Tried some storyboarding - Whose thinking and experience is influenced
> by whom? -http://robertblogs.wordpress.com/2010/01/31/software-craftsmanship-co...
... hadn't thought about folks entering negative influences! A valid
concern indeed. :) I'm all for creating a user experience which
allows for positive influences only. I feel that any such system
which compares folks in some way (influences, ranking, whatever) is
unlikely to succeed unless it's non-threatening and supportive.
I would be interested in contributing to a simple prototype.
Robert
On Feb 2, 2:28 am, James Martin <jimmymar...@gmail.com> wrote:
> On Mon, Feb 1, 2010 at 12:43 PM, Robert Reppel <rrep...@gmail.com> wrote:
> > Very interesting idea! What would be a measure of "collaborative
> > distance" among developers .... ?
>
> > Tried some storyboarding - Whose thinking and experience is influenced
> > by whom? -
> >http://robertblogs.wordpress.com/2010/01/31/software-craftsmanship-co...