Is it not reasonable to suppose that it was an idea before its time, and useless in the particular historical context and implementational form in which it appeared?
There has always been utility for mechanical computation, but it's entirely possible that the world simply did not have an application for The Analytical Engine in the 1830s-1840s because other sectors of technology and the economy simply hadn't evolved to a level where they could effectively utilise it, especially given its physical properties--its size, scale, and energy consumption.
I don't know that for a fact, and can't effectively gauge the merits of my own suggestion, as I am neither a mathematician nor a competent historian of the intellectual, scientific and commercial zeitgeist of that period. But, for the sake of argument, is it not possible that this invention fell into the "interesting, novel, but useless" category?
Now, as for the telephone:
1) From the point of view of the telegraph establishment, it was a competitor;
2) Unintelligible voice really is useless. They just weren't far-sighted enough to see that the voice quality could improve, and indeed, it was a quite a long time before it did. Local loop quality improved first. Long-distance toll voice really didn't begin to sound good until digital trunking came along. Ask your grandparents or great-grandparents what coast-to-coast long distance phone calls sounded like in the era of analog lines and waveguide-type multiplexing technology;
3) In the heyday of the telegraph era, deploying lines was an extremely expensive and capital-intensive process, and it wasn't until other technological advancements that made possible various multiplexing and aggregation schemes (frequency-division, and later digital TDM) came along that the idea of running a copper line into every home really got to be realistic[1]. I agree that Western Union was a bit shortsighted in turning down this patent, but one could hardly blame them for thinking that universal telephone service wasn't economically possible. That's like selling a business idea today that relies on everyone having a 10 terabit fiber cable run to their home. Yeah, it's possible, and I have no doubt someone will make fun of me in a decade or two for naysaying it in any way, but would you invest $2bn in a related patent today?
What mistake did all these very smart men make in common? They focused on what the technology could not do at the time rather than what it could do and might be able to do in the future.
I don't disagree, but that needs to be fleshed out. No viable entrepreneurial venture can succeed solely on the basis of what it is logically possible for the technology to someday do, or what it could, in principle, in theory, one day. There is a need to realise a return in a usefully short period of time that is also unanimously acceptable to a coterie of investors with varying needs in terms of payoff time frame and patience.
Thus, you need a practical plan for getting to point B, making the technology do X. Even the most far-fetched, high-risk, R&D-driven ventures entail a proposal to concretely deploy and commercialise technology in a period that is usefully short and politically palatable, and that means everyone involved is somewhat constrained to what can be practically envisaged in terms of today's possibilities. One can make some leaps of faith, some intelligent extrapolations and some prescient forecasts, but ultimately, it's something expressed largely in the observational language and ontology of today.
Thus, I can't bring myself to fault someone for doubting, in 1995, that the consumer web is going to be what it is today, or even be what it was six or seven years later, in the early 2000s. It was possible--perhaps even reasonable--to suppose so, but would you have bet the farm on it? Your retirement savings? I'm not sure I would have (not that I have a farm or retirement savings, pero bueno).
[1] http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=451163...