The "unscientific" side might be best explained by a little editorializing.
Theory, experiment and now computation have ended up split between specilaists because the required knowledge for each can no longer practically fit into one education. The effect of this is that since one brain has been separated into many heads, lots of partly-formed ideas must be communicated between them to replace what used to be internal musing.
Lots of published theory has little experimental or numerical footing; theory is the superego of the split mind and plays the closest role to imagining goals and setting courses.
Experiment would be lost without theory: experimental papers are relatively light on interpretation and usually serve to "check out" theorized results. To continue the metaphor they're the Id: sometimes experiments come up with results that no rational mind could have expected, grabbing the wheel and bringing theory crashing back down to earth, and upon reaching the earth again launching it off in amazing new directions.
Computational physicists (I'm including anyone who spends most of their time with numbers in this) are left as the integrating Ego, trying to write code that models the theories and analyses the data and on a good day gives the same answers for both.
Finally, I should add that all of these intermediate results are very carefully worded to be true, claiming no more than can be claimed. The system would still work if they were phrased assertively but there's such a strong culture of truth that nobody does that. So, you can trust the papers, but not so much the magazine write-ups which usually clobber the careful stepping.
Theory, experiment and now computation have ended up split between specilaists because the required knowledge for each can no longer practically fit into one education. The effect of this is that since one brain has been separated into many heads, lots of partly-formed ideas must be communicated between them to replace what used to be internal musing.
Lots of published theory has little experimental or numerical footing; theory is the superego of the split mind and plays the closest role to imagining goals and setting courses.
Experiment would be lost without theory: experimental papers are relatively light on interpretation and usually serve to "check out" theorized results. To continue the metaphor they're the Id: sometimes experiments come up with results that no rational mind could have expected, grabbing the wheel and bringing theory crashing back down to earth, and upon reaching the earth again launching it off in amazing new directions.
Computational physicists (I'm including anyone who spends most of their time with numbers in this) are left as the integrating Ego, trying to write code that models the theories and analyses the data and on a good day gives the same answers for both.
Finally, I should add that all of these intermediate results are very carefully worded to be true, claiming no more than can be claimed. The system would still work if they were phrased assertively but there's such a strong culture of truth that nobody does that. So, you can trust the papers, but not so much the magazine write-ups which usually clobber the careful stepping.